In view of the complex power quality disturbance sources and the difficult management of user-side power quality compensation equipment in power grid, this paper presents an algorithm for identifying the operation status of user-side compensation equipment and evaluating its effect based on the monitoring data of power quality on power grid side. Thus, the non-intrusive monitoring of compensation equipment on the user side can be realized by using the monitoring data. Firstly, the probabilistic neural network is used to classify the working conditions of compensation equipment by taking the voltage deviation, three-phase voltage unbalance, total harmonic distortion rate and long-term voltage flicker as the input parameters, so as to realize the evaluation of the operation status of the compensation equipment. Secondly, by comparing the results of power quality evaluation before and after the operation of compensation equipment, this paper uses the fuzzy comprehensive evaluation method to achieve the effect evaluation of compensation equipment. Taking the monitoring data of a substation in Anhui Power Grid as an example, the practicability of the proposed state identification method and the effect evaluation method of the compensation equipment is analyzed.