It is more difficult that cognizance and punishment of organized crime groups with development of the Internet. According to the characteristics of the organized crime groups, we are using co-offending networks analysis methods and data mining techniques to identifying organized crime structures and their constituent entities. An novel algorithm for mining organized crime groups is proposed. The goal of our work is to improve the efficiency of the organized crime detection for extracting information from large real-life crime datasets to obtain evidence of the organized crime group. Experimental results show that the algorithm of time performance is superior compared with other existing algorithms.