Accurate forecasting of wind and solar power generation data is a fundamental and crucial step towards promoting the efficient utilization and sustainable development of clean energy. In this context, this paper proposes a wind and solar power generation prediction model based on Wavelet Decomposition and CNN-GRU-LSTM-Attention. First, the raw wind and solar power generation data is decomposed into subsequences of different frequencies through wavelet decomposition, capturing the time-frequency features within the data. Next, Convolutional Neural Networks (CNN) are employed to extract local features of the time series, while Gated Recurrent Units (GRU) and Long Short-Term Memory (LSTM) networks are used to model the temporal dependencies of the data, with an attention mechanism introduced to dynamically assign weights to different time steps. Finally, the model is validated using wind and solar power generation data from a region in China in 2023. The results demonstrate that the proposed model leads to significant improvements across multiple evaluation metrics in the wind and solar power generation forecasting scenario. It not only greatly reduces errors but also significantly enhances the fitting performance.