针对传统安稳期判别方法在复杂风浪环境下，由于判别信息不充分、直升机起降状态难把握而导致的判别结果准确性和可靠性不足的现状，提出了一种多元混合模型，该模型结合了并行双向预测分支和直升机动力学约束链路，优化了环境信息的利用方式，将船舶运动和风场预测结果转化为直升机起降特征。模型基于P-BiLSTM框架，通过并行双向递推结构实现船舶运动包络与风速时程的同频预报，形成环境-运动信息融合预测框架，并通过动力学约束链路快速计算舰-机相对姿态和操纵状态，为舰载直升机起降作业提供准确状态信息。实测数据与仿真验证结果表明，模型在风浪环境下具有较高预测精度，可为起降辅助决策提供有效支持。

To address the limitations of traditional stability period discrimination methods, which suffer from insufficient discriminative information and difficulty in accurately identifying helicopter takeoff and landing states under complex wind and wave conditions, a multivariate hybrid model is proposed. This model integrates parallel bidirectional prediction branches with helicopter dynamics constraints, optimizing the utilization of environmental data. It transforms ship motion and wind field prediction results into helicopter takeoff and landing features. The model is built on a P-BiLSTM framework, employing a parallel bidirectional recursive structure to achieve synchronized forecasting of ship motion envelopes and wind speed time series. This forms an environment-motion information fusion prediction framework. Additionally, the dynamic constraint link rapidly computes the relative attitude and maneuvering state between the ship and helicopter, providing accurate state information for shipborne helicopter operations. Experimental data and simulation results demonstrate that the model achieves high prediction accuracy in wind and wave environments, offering effective support for takeoff and landing decision making.