Due to the characteristics of the mobile device, it is difficult to implement the existing network discovery algorithm based on clique in mobile social network. In order to solve the problem, through analyzing the mobile user behaviors, an algorithm of community discovery algorithm based on circuit merging is proposed. Firstly, the mobile social network is constructed according to the mobile user behaviors. The method finds the k-circuit as community core by using the k-Elementary Circuits (k-EC) finding algorithm. And then the community cores that have common nodes will be merged according to some given rules. Secondly, the remainder discrete nodes are added to elementary communities according to the mobile user correlation degree. Finally, the algorithm is proved more accurate and feasible in the experiments by using the public data sets and the simulated data sets.