<p>With the continuous increase in the penetration rate of renewable energy in the power system, its volatility and intermittency pose serious challenges to the stable operation of the power grid. This article proposes a power demand and renewable energy generation prediction model based on a dual stream TBATS-CNN-LSTM hybrid neural network for the complex interaction between the power system and renewable energy. This model combines the multi seasonal processing capability of TBATS model, the spatial feature extraction capability of convolutional neural network (CNN), and the temporal dependency modeling advantage of long short-term memory network (LSTM), which can effectively capture the complex changes in renewable energy generation and power load. By constructing a comprehensive simulation system covering wind power generation, photovoltaic power generation, and traditional power loads, the superiority of the model in terms of prediction accuracy and generalization ability was verified. The research results indicate that the model can provide effective technical support for optimizing the operation of the power system under high proportion of renewable energy access, and promote the large-scale consumption of renewable energy. This work has important theoretical significance and practical value for promoting the construction of new power systems and low-carbon transformation of energy structures.</p>