Objective Photovoltaic PV arrays have multiple peak points under complex operating conditions and traditional maximum power point tracking MPPT control algorithms cannot effectively track them. To address the problems of traditional algorithms such as being prone to getting trapped in local optimal values slow convergence speed and low tracking accuracy a PV multi-peak MPPT method based on a hybrid algorithm of improved grey wolf optimization IGWO and perturbation and observation PO is proposed. Methods First the GWO algorithm was used to initialize the population positions so as to obtain the initial maximum power. Then a global search was conducted by iteratively updating the positions of the grey wolves to guide them toward the maximum power point. Finally upon reaching the maximum iteration count or being in the vicinity of the maximum power point the algorithm switched to the variable-step perturbation and observation PO method. The perturbation step size was adaptively adjusted to enhance the local search ability of the system and accurately locate the maximum power point. Results Simulation and experimental results showed that the IGWO&PO algorithm did not fall into local optima. In terms of convergence speed the IGWO&PO algorithm was faster than the GWO algorithm and the hybrid algorithm of bacterial foraging optimization algorithm and particle swarm optimization BFOA-PSO . In terms of tracking accuracy the IGWO&PO algorithm improved by 4. 2% compared with the GWO algorithm. Conclusion The IGWO&PO algorithm effectively overcomes the tendency of traditional MPPT methods to become trapped in local optima. It achieves higher tracking accuracy and faster convergence while reducing oscillation amplitude during the optimization process thereby enhancing the overall energy utilization efficiency of PV systems.