摘要： The adaptive harmonic detection method based on the adaptive noise cancellation principle has the advantages of strong adaptability and good robustness. However, when the step size of the algorithm is selected, there is a conflict between the convergence speed and the steady-state accuracy. To solve this problem, this paper proposes an improved variable step size adaptive harmonic detection method. The algorithm uses the sliding integrator to find the feedback that can truly reflect the tracking situation, and substitutes it into the step size iteration formula based on L2 norm to achieve the step size adjustment. In order to further verify the superiority and practicability of the improved algorithm, the algorithm is applied to the photovoltaic-active power filter (PV-APF) system, and simulation verification is performed in MATLAB. The simulation results show that under the same simulation conditions, the convergence speed of the improved algorithm is half a cycle faster than that of the traditional method, and the steady-state error is smaller. In the application of PV-APF system, under normal light, the dual functions of grid-connected photovoltaic system and harmonic compensation are realized. The total harmonic distortion (THD) of the system is 2.09%, which effectively improves the power quality of photovoltaic system.