简单的答案：浮动使用有限精度的二进制编码来表示数字，所以各种操作都会丢失一些精度。

***页面有很多信息（可能太多）。

另见：

Possible Duplicate:
Python float - str - float weirdness

Python float division does not appear to have accurate results. Can someone explain why?

>>>3.0/5
0.59999999999999998

Short answer: Floats use finite-precision binary encoding to represent numbers, so various operations lose some precision.

The Wikipedia page has a lot of information (maybe too much).

See also: How do I use accurate float arithmetic in Python?