为便于阅读而重写：

# 2D array/ matrix

# 5 rows, 5 cols
rows_count = 5
cols_count = 5

# create
#     creation looks reverse
#     create an array of "cols_count" cols, for each of the "rows_count" rows
#        all elements are initialized to 0
two_d_array = [[0 for j in range(cols_count)] for i in range(rows_count)]

# index is from 0 to 4
#     for both rows & cols
#     since 5 rows, 5 cols

# use
two_d_array[0][0] = 1
print two_d_array[0][0]  # prints 1   # 1st row, 1st col (top-left element of matrix)

two_d_array[1][0] = 2
print two_d_array[1][0]  # prints 2   # 2nd row, 1st col

two_d_array[1][4] = 3
print two_d_array[1][4]  # prints 3   # 2nd row, last col

two_d_array[4][4] = 4
print two_d_array[4][4]  # prints 4   # last row, last col (right, bottom element of matrix)

如果你真的想要一个矩阵，你最好使用numpy。numpy中的矩阵运算通常使用二维数组类型。创建新阵列有多种方法；其中最有用的是zeros函数，它接受一个形状参数并返回一个给定形状的数组，值初始化为零：

>>> import numpy
>>> numpy.zeros((5, 5))
array([[ 0.,  0.,  0.,  0.,  0.],
       [ 0.,  0.,  0.,  0.,  0.],
       [ 0.,  0.,  0.,  0.,  0.],
       [ 0.,  0.,  0.,  0.,  0.],
       [ 0.,  0.,  0.,  0.,  0.]])

以下是创建二维数组和矩阵的一些其他方法（为了紧凑，去掉了输出）：

numpy.arange(25).reshape((5, 5))         # create a 1-d range and reshape
numpy.array(range(25)).reshape((5, 5))   # pass a Python range and reshape
numpy.array([5] * 25).reshape((5, 5))    # pass a Python list and reshape
numpy.empty((5, 5))                      # allocate, but don't initialize
numpy.ones((5, 5))                       # initialize with ones

numpy也提供了一种矩阵类型，但它不再推荐用于任何用途，将来可能会从numpy中删除。

为便于阅读而重写：

# 2D array/ matrix

# 5 rows, 5 cols
rows_count = 5
cols_count = 5

# create
#     creation looks reverse
#     create an array of "cols_count" cols, for each of the "rows_count" rows
#        all elements are initialized to 0
two_d_array = [[0 for j in range(cols_count)] for i in range(rows_count)]

# index is from 0 to 4
#     for both rows & cols
#     since 5 rows, 5 cols

# use
two_d_array[0][0] = 1
print two_d_array[0][0]  # prints 1   # 1st row, 1st col (top-left element of matrix)

two_d_array[1][0] = 2
print two_d_array[1][0]  # prints 2   # 2nd row, 1st col

two_d_array[1][4] = 3
print two_d_array[1][4]  # prints 3   # 2nd row, last col

two_d_array[4][4] = 4
print two_d_array[4][4]  # prints 4   # last row, last col (right, bottom element of matrix)

公认的答案是正确的，但我花了一段时间才明白，我也可以使用它创建一个完全空的数组。

l =  [[] for _ in range(3)]

结果

[[], [], []]

在Python中，您将创建一个列表列表。您不必提前声明维度，但可以这样做。例如：

matrix = []
matrix.append([])
matrix.append([])
matrix[0].append(2)
matrix[1].append(3)

现在矩阵[0][0]==2，矩阵[1][0]==3。您还可以使用列表理解语法。本示例使用它两次来构建“二维列表”：

from itertools import count, takewhile
matrix = [[i for i in takewhile(lambda j: j < (k+1) * 10, count(k*10))] for k in range(10)]

试试看：

rows = int(input('Enter rows\n'))
my_list = []
for i in range(rows):
    my_list.append(list(map(int, input().split())))