It affects the output of photovoltaic power by meteorological factors such as irradiance and temperature

showing significant randomness and volatility

which poses difficulties for the high-precision prediction of photovoltaic power.It proposes a CNN-LSTM-Attention hybrid prediction model optimized based on the improved sparrow search algorithm in response to the above problems.This model utilizes a convolutional neural network to extract the local spatial features of the input data

combines a long short-term memory network to capture the long-term temporal dependence rules of the power sequence

and introduces an attention mechanism to dynamically weight the hidden states

thereby enhancing the model's perception ability of features at critical moments.According to the local convergence problem existing in the standard sparrow search algorithm

through integrating Tent chaotic mapping

nonlinear adaptive weights and Levy flight strategies

it automatically optimizes based on the hyperparameters of the hybrid model.The simulation results based on the measured data of a certain photovoltaic power station show that the model can effectively reduce the prediction deviation caused by random meteorological factors

and the prediction output is in high agreement with the actual observed values

which has certain application value in the field of photovoltaic power prediction.