The successful landing of carrier-based aircraft is a core aspect of the aircraft carrier combat system, directly affecting the combat effectiveness of the carrier group. This paper focuses on the intelligent assessment of land-based landing techniques for carrier-based aircraft, proposing an innovative solution—a deep learning evaluation algorithm based on a CNN-LSTM network. Compared to independent deep learning models, the CNN-LSTM hybrid architecture demonstrates superior spatiotemporal feature extraction capabilities. This algorithm efficiently constructs an accurate deep learning assessment model by deeply mining flight data during the landing process and integrating feedback from the landing signal officer (LSO). The experimental results show that the algorithm performs excellently in terms of accuracy, stability, and real-time performance, overcoming the limitations of traditional evaluation methods that rely on subjective expert scoring or overly simplified parameter thresholds, providing strong support for the enhancement of pilot skills. At the same time, this breakthrough makes data-driven reporting, predictive maintenance alerts for landing gear systems, and AI-assisted LSO training modules possible.