Dam deformation is affected by many factors, and its abnormal observed value is not surely the gross error. In order to effectively identify the gross error in the safety monitoring data of dam, the statistical model used in safety monitoring of dam and its bases is introduced into the gross error recognition of monitoring data on the basis of analyzing the statistical model theory and the reason that the gross errors are generated. First of all, the data containing the gross error are used to establish the statistical model, and then according to the residual error between the fitting results of statistical model and the real measured value, the quartile method is used to set threshold and recognize the gross errors. For some concrete gravity dam, after the gross errors are added into the monitoring data of tension wire on the top of dam, the actual situation is simulated. Through this method, the added gross errors are completely recognized.