In recent years, the rapid development of geospatial intelligence and remote sensing technologies has brought new ways to study land cover. By combining different data sources, such as hyperspectral and LiDAR data, the AI model's land cover classification accuracy reaches new heights. Yet, due to the difference in data extraction and processing, the fusion of two data types is challenging. This paper presents a novel model that combines two input data sources, hyperspectral and LiDAR, to improve the application of machine learning (ML) or artificial intelligence (AI) in land cover classification. This is a new network based on GNNs and Mamba, leveraging multi-source HSI and LiDAR data to enhance classification results. In this paper, the Gated Recurrent Units (GRUs) and Vision Transformers (ViTs) were selected to enhance performance. By integrating GRUs and ViTs into GNN and Mamba frameworks, the proposed networks aim to leverage the strengths of these components to address challenges in multi-source HSI and LiDAR data classification. All 3 models show outstanding performance across the 3 datasets (MUFFL, Trento, and Houston). By introducing cutting-edge and diverse ML/AI models and components tailored to different tasks, this paper aims to explore the application prospects of ML/AI in land cover studies that could benefit the wider community.