There are a large number of outliers in the PV array operation data. The outlier will bring difficulties to the functions such as PV array performance analysis, modeling, and fault diagnosis. In order to effectively clean the outlier in the PV array operation data, this paper proposes a cleaning method for PV array outlier based on moving standard deviation. It analyzes the source and distribution characteristics of the array outlier data and proposes the algorithm based on moving standard deviation. The curve’s rising of the sliding standard deviation set is used as the basis for judging the outlier data. Finally, through the case analysis and comparison of quartile method, the results show that the algorithm can effectively reduce the cleaning error caused by the concentration distribution of the outlier. This work is supported by National Natural Science Foundation of China (No. U1765104) and Science and Technology Project of State Grid Hebei Electric Power Co., Ltd. (No. 5204BB170007) Research and Application of Regional Landscape Resources and Mid-long Term Forecast Technique for Power Generation Ability.