Cognitive radio networks are a new technology based on which unlicensed users are allowed access to licensed spectrum under the condition that the interference perceived by licensed users is minimal. That means unlicensed users need to learn from environmental changes and to take appropriate decisions regarding the access to the radio channel. This is a process that can be done by unlicensed users in a cooperative or non-cooperative way. Whereas the non-cooperative algorithms are risky with regard to performance, the cooperative algorithms have the capability to provide better performance. Our paper reports therefore a new fuzzy logic based decision making algorithm for channel selection. The underlying decision criterion considers statistics of licensed users channel occupancy as well as information about the competition level of unlicensed users. Our theoretical studies indicate that the unlicensed users can obtain an efficient sharing of the available channels. Simulation results are reported to demonstrate the performance and effectiveness of the suggested algorithm.