Provided is a system detection method and system based on data slice consistency. The method comprises the steps that a user transaction behavior trace is obtained, a user control flow model is established according to the user transaction behavior trace, and a user data flow model is established according to user transaction data information; according to the control flow model and the data flow model, a user behavior model is obtained through integration; according to user behaviors and data characteristics, slices of the behavior model and the data model are analyzed to obtain a dynamic behavior slice and a data slice, and an overall slice of a user transaction system including the user behaviors and the data characteristics is analyzed to obtain data dependence data; according to the data dependence data, the data consistency of the behavior slice is detected to obtain a consistency detection result; according to the consistency detection result, the user behaviors and data anomaly are detected, and transaction system vulnerable zones causing data anomaly are locked.