如.NET Framework文档中所述

我们不建议你将数组列表类用于new 发展。相反，我们建议您使用通用的List<T> 类。ArrayList类被设计用来保存异构类 对象的集合。然而，它并不总是提供最好的 的性能。相反，我们建议如下: 对于异构对象集合，使用List<Object>(在c#中)或List(of Object)(在Visual Basic中)类型。 对于对象的同构集合，使用List<T>类。

请参见不应使用非泛型集合

ArrayList不是类型安全的，而List<T>是类型安全的。简单:)。

我认为，ArrayList和List<T>之间的区别是:

List<T>, where T is value-type is faster than ArrayList. This is because List<T> avoids boxing/unboxing (where T is value-type). Many sources say - usually ArrayList used just for backward compatibility. (is not a real difference, but i think it is important note). Reflection is easier with nongeneric ArrayList then List<T> ArrayList has IsSynchronized property. So, It is easy to create and use syncronised ArrayList. I didin't found IsSynchronized property for List<T>. Also Keep in mind this type of synchronization is relatively inefficient, msdn): var arraylist = new ArrayList(); var arrayListSyncronized = ArrayList.Synchronized(arraylist Console.WriteLine($"syncronized {arraylist.IsSynchronized}"); Console.WriteLine($"syncronized {arrayListSyncronized.IsSynchronized}"); var list = new List<object>(); var listSyncronized = ArrayList.Synchronized(list); Console.WriteLine($"syncronized {list.IsSynchronized}");//error, no such prop Console.WriteLine($"syncronized {list.IsSynchronized}");//error, no such prop ArrayList has ArrayList.SyncRoot property which can be used for syncronisation (msdn). List<T> hasn't SyncRoot property, so in the following construction you need to use some object if you use List<T>: ArrayList myCollection = new ArrayList(); lock(myCollection.SyncRoot) // ofcourse you can use another object for this goal { foreach (object item in myCollection) { // ... } }

使用List<T>可以防止强制转换错误。它对于避免运行时强制转换错误非常有用。

例子:

在这里(使用ArrayList)，您可以编译这段代码，但稍后会看到一个执行错误。

ArrayList array1 = new ArrayList();
array1.Add(1);
array1.Add("Pony"); //No error at compile process
int total = 0;
foreach (int num in array1)
{
 total += num; //-->Runtime Error
}

如果使用List，可以避免以下错误:

List<int> list1 = new List<int>();
list1.Add(1);
//list1.Add("Pony"); //<-- Error at compile process
int total = 0;
foreach (int num in list1 )
{
 total += num;
}

参考: MSDN

再加上以上几点。在64位操作系统中使用ArrayList占用的内存是32位操作系统的2倍。同时，泛型列表list <T>占用的内存比ArrayList少得多。

例如，如果我们在32位中使用一个19MB的数组列表，那么在64位中它将占用39MB。但是如果你有一个32位的8MB的通用列表list <int>，它在64位只需要8.1MB，这与ArrayList相比相差了481%。

来源:数组列表与通用列表的基本类型和64位

To me its all about knowing your data. If I am continuing to expand my code on the basis of efficiency, I would have to choose the List option as a way of deciphering of my data w/o the unnecessary step of always wondering about types, especially 'Custom Types'. If the machine understands the difference and can determine on it's on what type of data I'm actually dealing with then why should I get in the way and waste time going thru the gyrations of 'IF THEN ELSE' determinations? My philosophy is to let the machine work for me instead of me working on the machine? Knowing the unique differences of different object code commands goes a long way in making your code as efficient.

汤姆•约翰逊 (一项…一个出口)