Abstract: Perturbation and Observation Method (P&O), also known as Hill Climbing Method, is a commonly used self optimizing control strategy in photovoltaic power generation systems to achieve the maximum power point of solar energy perturbation caused by photovoltaic panels. The conventional disturbance observation method cannot quickly track the maximum power point as the external radiation intensity and temperature constantly change, and there is significant oscillation when approaching the maximum power point. Therefore, this article proposes an improved disturbance observation method to overcome oscillations near the maximum power point and improve tracking response speed. Firstly, based on the original one-step power prediction, a two-step power prediction is adopted; Secondly, an algorithm for optimizing the duty cycle is adopted, which multiplies the rate of change of the duty cycle by a coefficient that is proportional to the voltage of the photovoltaic array and inversely proportional to the current. The step size of the improved algorithm decreases as it approaches the maximum power point; Meanwhile, use a DC-DC Boost circuit powered by photovoltaic array modules to test the proposed concept; Finally, a photovoltaic power generation system model was built in Matlab/Simulink, and the simulation results showed that the efficiency of the algorithm at the maximum power point was about 3.2% higher than the traditional perturbation observation method (P&O) algorithm and conductance increment (INC) algorithm, further proving the feasibility and correctness of the improved control strategy.