This paper introduces a hybrid deep learning model for maneuver decision-making in air combat scenarios. The model combines a stacked sparse autoencoder (SSAE) for dimensionality reduction of high-dimensional, dynamic time-series combat data with a long short-term memory (LSTM) network to model the quantitative relationship between the reduced data and maneuver control variables. Key features of the model include leveraging time-series data as the foundation for decision-making, which aligns closely with real-world processes, and using SSAE to enhance prediction accuracy by reducing data dimensionality. Additionally, the model outputs maneuver control variables, enabling flexible and effective control of maneuvers. Experimental results demonstrate that the proposed approach significantly improves prediction accuracy and convergence speed, making it a robust solution for autonomous air combat decision-making systems.