A chaotic adaptive weight improved snake optimization algorithm is proposed to solve the issues of slow convergence speed, low convergence accuracy, and susceptibility to local optima in the snake optimization algorithm of single model parameter extraction of photovoltaic cells. The initialization of chaos and the setting of adaptive weights will dynamically better allocate the proportion of search of food, combat, and mating in snake algorithm. By comparing theoretical and experimental voltametric data under different light intensities and temperatures, the improved algorithm is verified. Compared to snake optimization algorithm, the results show that the improved algorithm improves convergence speed by about 219.7%, accuracy by about 58.40%, and stability by about 49.57%. Finally, the influence of light intensity on photogenerated current is found to be greater than that of temperature, and the influence of temperature on reverse saturation current, series resistance, parallel resistance, and ideal factor is greater than that of light intensity.