Edit: drfyer, you beat me.

\"Fuzzy logic\" is usually referring to a way of creating results not based on an equation or even a set of equations, but to base it on a set of \"rules\". These rules are defined like real-world rules.
A simple example is the classification of tomatoes based on color:
green - red - dark red
corresponds to ripeness
not ripe - ripe - overly ripe
seems simple. However, the rules are not either-or, nor are they mutually exclusive. Therefore,
- every rule may or may not be met _simultaneously_, to a varying degree, as defined in the rule truth and
- the result may take into account the different rules to varying degrees, maybe also depending on the actual fulfillment of the rules.
The tomato might be \"somewhat green\", i.e. 20% green, and also \"mostly red\" (75% red), and maybe even \"slightly dark\" (5% dark red). The system would maybe evaluate the ripeness of this tomato to \"ripe\", with a sureness of maybe 95%.
This still looks like simple number comparison, but due to the rule bases it\'s much more easy to define \"shades\", both in terms of output and input; it\'s more easy to define what\'s going to happen \"inbetween\" the usual states.
You can compare it to a large set of input vs. output tables that get combined by another set of tables and so on.
This is most likely the worst explanation of this subject that has ever been given, though.
Therefore, here is a better explanation to be downoaded (pdf format):
http://www.minco.com/site/wordpdf/pdf00201-fuzzy.pdf