The invention relates to the technical field of industrial information collection and analysis, and aims to provide a predictive industrial equipment maintenance method and maintenance system based onedge computing. The invention discloses a predictive industrial equipment maintenance method based on edge calculation. The method comprises the following steps: S1, collecting industrial data; S2, preprocessing the industrial data to obtain standard industrial data; S3, performing feature selection to obtain a final feature; S4, constructing an abnormity detection model, inputting the standard industrial data and the final features into the abnormity detection model, and carrying out abnormity detection on the standard industrial data; and S5, constructing a fault prediction model, judging whether the industrial equipment corresponding to the standard industrial data has a fault or not, if yes, issuing early warning information, and if not, entering the step S1. A maintenance system comprises industrial equipment, an edge gateway and a server which are sequentially connected. Predictive maintenance of industrial equipment can be achieved, invalid manual operation is avoided, and automatic operation is facilitated.