New data methods for forecasting systems of regional sea level prediction based on climate variables using both classical time series models and modern machine learning methods will be presented. Sea level rise induced by climate change poses a significant threat to coastal cities, many of which serve as major economic hubs due to their strategic coastal locations, such as New York and Shanghai. To mitigate the increasing risks, accurate predictive models are urgently required to assess potential impacts effectively. In this study, we treat sea level data as signal data, leveraging its temporal structure to apply advanced signal processing techniques. Specifically, we employ Seasonal-Trend Decomposition using Loess (STL) to isolate the underlying trend by removing seasonal components and noise. This extracted trend is then used to train predictive models. We propose a hybrid framework that integrates STL decomposition with deep learning architectures, focusing on CNN-LSTM and CNN-GRU，CNN-Mamba networks and physics infromation to capture both spatial and temporal dependencies in sea level data. 

提出一种基于气候变量、结合经典时间序列模型与现代机器学习方法的区域海平面预测新型数据方法体系。由气候变化引起的海平面上升对沿海城市构成了重大威胁，而许多沿海城市由于其战略性的地理位置同时也是重要的经济中心，例如纽约和上海。为了降低不断增加的风险，迫切需要构建高精度的预测模型，以有效评估潜在影响。在本研究中，我们将海平面数据视为一种信号数据，利用其时间结构特性来应用先进的信号处理技术。具体而言，我们采用基于局部加权回归的季节–趋势分解方法（Seasonal-Trend Decomposition using Loess, STL），通过去除季节性成分和噪声来提取潜在趋势。随后利用该趋势序列训练预测模型。我们提出了一种混合框架，将 STL 分解方法与深度学习架构相结合，重点采用 CNN-LSTM、CNN-GRU 以及 CNN-Mamba 网络，并融合物理信息约束，以同时捕捉海平面数据中的空间与时间依赖关系。

海平面升降,深度学习,人工智能,物理信息约束