As a result of changes in approach from traditional to virtual banking system, security in data exchange has become more important; thus, it seems essentially necessary to present a pattern based on smart models in order to reduce fraud in this field. A new algorithm has been provided in this article to improve security and to specify the limits of giving special services to internet banking users in order to pave appropriate ground for virtual banking. In addition to identifying behavioral models of customers, this algorithm compares the behaviors of any customer with this model and finally computes the rate of trust in customer’s behavior. The hybrid data-mining and knowledge based structure has been adapted in this algorithm according to fuzzy systems. In this research, qualitative data gathered from interviews with banking experts, analyzed by Expert Choice to identify the most important variables of customer behavior analysis, and to analyze customer behavior, customer bank Internet transaction data for a period of one year by MATLAB and Clementine. The results of this survey indicate that the potential of the given structure to recognize the rate of trust in internet bank user’s behavior might be at reasonable level for experts in this area.

Keywords