The P-U curve of photovoltaic system has a multi-peak characteristic under partial shading condition. In this case

the conventional algorithm is easy to fall into local optimum when tracking the maximum power point

and the intelligent optimization algorithm takes a long time to track. In this regard

this paper designs a three-step composite maximum power point tracking(MPPT) algorithm combining the improved perturbation and observation(IPO) algorithm with the improved beetle swarm optimization(IBSO) algorithm. The algorithm uses IPO algorithm to quickly track to the maximum power point at first

and uses the power value of this point to adjust the voltage search range. Then

it uses the IBSO algorithm to perform global optimization in the range to ensure the search accuracy. Finally

after the IBSO algorithm tracks near the maximum power point

the IPO algorithm is switched again to speed up the tracking speed and reduce power oscillation. Simulation comparison between the three-step algorithm and the IPO

BSO and CSA-IPO algorithm is carried out. The results show that

the proposed three-step algorithm is not easy to fall into the local optimum

moreover

it can track the maximum power point accurately and quickly

and has smaller power oscillation during the tracking process.