In case of large amount of the search engine based applications, mobile e-commerce has established a more interest under both industry and academia. From that mining the user behavior and prediction of the user to analysis the mobile commerce behaviors based on their actions are most important. To perform these steps, previous work proposed a novel structure called Mobile Commerce Explorer (MCE). It can be performed in three ways 1) Similarity Inference Model (SIM) for measuring the similarities amongst stores and items 2) Personal Mobile Commerce Pattern Mine (PMCP-Mine) algorithm for well-organized discovery of mobile users Personal Mobile Commerce Patterns (PMCPs); and 3) Mobile Commerce Behavior Predictor (MCBP) for prediction of possible mobile user behaviors. Assigning the weight values for each item in the mobile transaction database finds the best frequent pattern mining from the mobile user pattern .Proposed system considering the different weight values for each item and the select the best weight values to frequent pattern mining .Selection of best weight values from the different weight values we use particle swarm optimization algorithm system is initialized with a population of random solution such as different weight values and search for best weight values by updating invention. The particle swarm optimization changing the velocity of each particle toward finds best weight values at both local and global locations. After finding the weight values than derive the frequent pattern mining results from the existing transaction behavior of each mobile users .Weighted frequent pattern mining with PSO achieve an wide-ranging investigational estimation by replication and show that better accuracy outcome.