The selection of Frequency Hopping (FH) sequences is a crucial factor affecting the quality of FH communication in asynchronously organized networks. According to the requirement of FH communication in asynchronously organized networks, some definitions such as the frequency similarity, the sequence similarity, and the collective similarity are put forward, and then the similarity matrix of available FH sequences is constructed. The hierarchy clustering analysis method is introduced to analyze the clustering structure of FH sequences in characteristic space. Based on the principle that the similarity of suitable FH sequences used in different networks should be the least, the optimized sequences can be selected from different clusters finally.