The invention discloses a photovoltaic array maximum power point tracking method based on self-adaption drosophila melanogaster searching. When a photovoltaic array works, the shielding phenomenon is common, and the P-V output characteristics of the shielded photovoltaic array have multi-leak characteristics, so that the maximum power point tracking is likely to be trapped into a local extreme point. By adopting the combination of self-adaption drosophila melanogaster searching and a perturbation and observation method, photovoltaic array maximum power point tracking on the shielded condition can be achieved. Firstly, a self-adaption drosophila melanogaster algorithm is adopted, a global maximum power point is searched for, then, a small step perturbation and observation method is adopted for monitoring whether the external environment changes or not, if the external environment changes, drosophila melanogaster searching is executed again for performing maximum power point tracking, and if not, the existing output power is kept unchanged. According to the maximum power point tracking method, the novel biological group intelligent algorithm is adopted, global optimization is achieved, and the phenomena that a traditional conductance increment method, a perturbation and observation method and other methods can only perform tracking on the maximum power point on the condition that no shielding is performed and are likely to be trapped into a local extreme value, and when the external environment changes suddenly, misjudgment is likely to be caused are avoided. The method can effectively improve the maximum power point tracking stability and output efficiency of a photovoltaic system.