Due to the complexity of crop photosynthesis, a single date-driven photosynthetic rate model is difficult to accurately reflect its actual changes, resulting in errors between the model output and the actual value, which is difficult to meet the needs of optimal control of greenhouse environment considering crop growth information. Therefore, a modeling method based on error compensation ELM is proposed to establish the photosynthetic rate model. In this method, on this basis of the initial model established by extreme learning machine (ELM), the Gaussian mixture model (GMM) is used to describe the error mean of photosynthetic rate model, and then realize the error compensation of the initial model. Finally, simulation results show the effectiveness of the proposed modeling method.