这里有一个线程安全的方法来做到这一点:

public static class EnumerableExtension
{
    private static Random globalRng = new Random();

    [ThreadStatic]
    private static Random _rng;

    private static Random rng 
    {
        get
        {
            if (_rng == null)
            {
                int seed;
                lock (globalRng)
                {
                    seed = globalRng.Next();
                }
                _rng = new Random(seed);
             }
             return _rng;
         }
    }

    public static IEnumerable<T> Shuffle<T>(this IEnumerable<T> items)
    {
        return items.OrderBy (i => rng.Next());
    }
}

IEnumerable扩展方法:

public static IEnumerable<T> Randomize<T>(this IEnumerable<T> source)
{
    Random rnd = new Random();
    return source.OrderBy<T, int>((item) => rnd.Next());
}

使用基于Fisher-Yates Shuffle的扩展方法Shuffle任意(I)List:

private static Random rng = new Random();  

public static void Shuffle<T>(this IList<T> list)  
{  
    int n = list.Count;  
    while (n > 1) {  
        n--;  
        int k = rng.Next(n + 1);  
        T value = list[k];  
        list[k] = list[n];  
        list[n] = value;  
    }  
}

用法:

List<Product> products = GetProducts();
products.Shuffle();

上面的代码使用了备受批评的系统。选择交换候选的随机方法。它速度很快，但并不像它应该的那样随机。如果你需要更好的随机质量，可以使用System.Security.Cryptography中的随机数生成器，如下所示:

using System.Security.Cryptography;
...
public static void Shuffle<T>(this IList<T> list)
{
    RNGCryptoServiceProvider provider = new RNGCryptoServiceProvider();
    int n = list.Count;
    while (n > 1)
    {
        byte[] box = new byte[1];
        do provider.GetBytes(box);
        while (!(box[0] < n * (Byte.MaxValue / n)));
        int k = (box[0] % n);
        n--;
        T value = list[k];
        list[k] = list[n];
        list[n] = value;
    }
}

一个简单的比较可以在这个博客(WayBack Machine)上找到。

Edit: Since writing this answer a couple years back, many people have commented or written to me, to point out the big silly flaw in my comparison. They are of course right. There's nothing wrong with System.Random if it's used in the way it was intended. In my first example above, I instantiate the rng variable inside of the Shuffle method, which is asking for trouble if the method is going to be called repeatedly. Below is a fixed, full example based on a really useful comment received today from @weston here on SO.

Program.cs:

using System;
using System.Collections.Generic;
using System.Threading;

namespace SimpleLottery
{
  class Program
  {
    private static void Main(string[] args)
    {
      var numbers = new List<int>(Enumerable.Range(1, 75));
      numbers.Shuffle();
      Console.WriteLine("The winning numbers are: {0}", string.Join(",  ", numbers.GetRange(0, 5)));
    }
  }

  public static class ThreadSafeRandom
  {
      [ThreadStatic] private static Random Local;

      public static Random ThisThreadsRandom
      {
          get { return Local ?? (Local = new Random(unchecked(Environment.TickCount * 31 + Thread.CurrentThread.ManagedThreadId))); }
      }
  }

  static class MyExtensions
  {
    public static void Shuffle<T>(this IList<T> list)
    {
      int n = list.Count;
      while (n > 1)
      {
        n--;
        int k = ThreadSafeRandom.ThisThreadsRandom.Next(n + 1);
        T value = list[k];
        list[k] = list[n];
        list[n] = value;
      }
    }
  }
}

I'm bit surprised by all the clunky versions of this simple algorithm here. Fisher-Yates (or Knuth shuffle) is bit tricky but very compact. Why is it tricky? Because your need to pay attention to whether your random number generator r(a,b) returns value where b is inclusive or exclusive. I've also edited Wikipedia description so people don't blindly follow pseudocode there and create hard to detect bugs. For .Net, Random.Next(a,b) returns number exclusive of b so without further ado, here's how it can be implemented in C#/.Net:

public static void Shuffle<T>(this IList<T> list, Random rnd)
{
    for(var i=list.Count; i > 0; i--)
        list.Swap(0, rnd.Next(0, i));
}

public static void Swap<T>(this IList<T> list, int i, int j)
{
    var temp = list[i];
    list[i] = list[j];
    list[j] = temp;
}

试试这段代码。

你的问题是如何随机化一个列表。这意味着:

所有独特的组合都应该是可能发生的 所有唯一的组合应该出现在相同的分布(也就是无偏倚)。

由于“随机”，这个问题的大量答案不满足上述两个要求。

下面是一个紧凑的、无偏倚的伪随机函数，遵循Fisher-Yates shuffle方法。

public static void Shuffle<T>(this IList<T> list, Random rnd)
{
    for (var i = list.Count-1; i > 0; i--)
    {
        var randomIndex = rnd.Next(i + 1); //maxValue (i + 1) is EXCLUSIVE
        list.Swap(i, randomIndex); 
    }
}

public static void Swap<T>(this IList<T> list, int indexA, int indexB)
{
   var temp = list[indexA];
   list[indexA] = list[indexB];
   list[indexB] = temp;
}

这里是Fisher-Yates shuffle的实现，允许指定返回的元素数量;因此，在获取所需数量的元素之前，没有必要首先对整个集合进行排序。

交换元素的顺序与默认值相反;从第一个元素到最后一个元素，因此检索集合的一个子集与洗牌整个集合产生相同的(部分)序列:

collection.TakeRandom(5).SequenceEqual(collection.Shuffle().Take(5)); // true

该算法基于Durstenfeld在维基百科上的(现代)Fisher-Yates shuffle。

public static IList<T> TakeRandom<T>(this IEnumerable<T> collection, int count, Random random) => shuffle(collection, count, random);
public static IList<T> Shuffle<T>(this IEnumerable<T> collection, Random random) => shuffle(collection, null, random);
private static IList<T> shuffle<T>(IEnumerable<T> collection, int? take, Random random)
{
    var a = collection.ToArray();
    var n = a.Length;
    if (take <= 0 || take > n) throw new ArgumentException("Invalid number of elements to return.");
    var end = take ?? n;
    for (int i = 0; i < end; i++)
    {
        var j = random.Next(i, n);
        (a[i], a[j]) = (a[j], a[i]);
    }

    if (take.HasValue) return new ArraySegment<T>(a, 0, take.Value);
    return a;
}