Abstract: Accurate identification of solar cell model parameters has a great impact on PV module power prediction and maximum power point tracking,and it must be ensured to have high accuracy. Traditional intelligent algorithms can achieve a certain degree of parameter identification,but they all suffer from the problems of insufficient accuracy,slow convergence,and easy to fall into local optimality. To address such problems,a solar cell model parameter identification method based on the improved pelican optimization algorithm（IPOA）is proposed. In this algorithm,the population individuals are closely connected,and the position is updated by mutual learning of randomness,which has better effect than the traditional algorithm in engineering applications. At the same time,for the characteristics of this algorithm,a position updating strategy based on Jaya algorithm is introduced to make the candidate solutions of the population more optimal;the decreasing factor is improved to make the model better in the later stage of the iteration. The Lévy flight strategy is added,which effectively improves the algorithm accuracy. IPOA has good results under different solar irradiance,and the discrimination results fit well with the actual curves,indicating that IPOA can accurately and effectively identify the solar cell model parameters in different environments.