Abstract: Based on input/output (I/O) data of an unknown system,a new method for constructing a fuzzy system adaptively,which meets the accuracy requirements of the system,is put forward. To optimize the system rules,a new algorithm based on the binary search algorithm is designed,which can obtain an upper limit of rules to meet the accuracy requirements. Subsequently,the complexity of the algorithm is analyzed from the perspective of the algorithm running time,and the formula for estimating running time of the algorithm is proposed. The simulation data shows that the method is feasible and effective.