HSIMamba is a novel framework for hyperspectral image (HSI) classification that integrates bidirectional state space modules with convolutional neural networks (CNNs) to efficiently extract spectral features. The model uses bidirectional reversed convolutional neural network pathways to enhance spectral feature extraction and incorporates a specialized block for spatial analysis. By combining the efficiency of CNNs with the dynamic feature extraction capabilities of attention mechanisms, HSIMamba achieves high classification accuracy while reducing computational demands. It processes data bidirectionally, improving spectral feature extraction and integrating spatial information for comprehensive analysis. HSIMamba was tested on three widely recognized datasets—Houston 2013, Indian Pines, and Pavia University—and demonstrated exceptional performance, surpassing existing state-of-the-art models in HSI classification. The model's efficiency makes it suitable for deployment in scenarios with limited computational resources. HSIMamba redefines the standards of efficiency and accuracy in HSI classification, enhancing remote sensing applications. Hyperspectral imaging is crucial for environmental surveillance, agriculture, and other fields requiring detailed Earth surface analysis. The model's architecture includes a bidirectional processing mechanism and a spatial processing block, enabling efficient feature extraction and classification. HSIMamba outperforms traditional transformer-based models in terms of computational efficiency and classification accuracy, with reduced parameter count and FLOPs. The model's efficiency is validated through experiments on the three datasets, showing superior performance in overall accuracy, average accuracy, and Kappa coefficient. HSIMamba's efficiency and performance make it a promising candidate for hyperspectral image analysis in remote sensing.HSIMamba is a novel framework for hyperspectral image (HSI) classification that integrates bidirectional state space modules with convolutional neural networks (CNNs) to efficiently extract spectral features. The model uses bidirectional reversed convolutional neural network pathways to enhance spectral feature extraction and incorporates a specialized block for spatial analysis. By combining the efficiency of CNNs with the dynamic feature extraction capabilities of attention mechanisms, HSIMamba achieves high classification accuracy while reducing computational demands. It processes data bidirectionally, improving spectral feature extraction and integrating spatial information for comprehensive analysis. HSIMamba was tested on three widely recognized datasets—Houston 2013, Indian Pines, and Pavia University—and demonstrated exceptional performance, surpassing existing state-of-the-art models in HSI classification. The model's efficiency makes it suitable for deployment in scenarios with limited computational resources. HSIMamba redefines the standards of efficiency and accuracy in HSI classification, enhancing remote sensing applications. Hyperspectral imaging is crucial for environmental surveillance, agriculture, and other fields requiring detailed Earth surface analysis. The model's architecture includes a bidirectional processing mechanism and a spatial processing block, enabling efficient feature extraction and classification. HSIMamba outperforms traditional transformer-based models in terms of computational efficiency and classification accuracy, with reduced parameter count and FLOPs. The model's efficiency is validated through experiments on the three datasets, showing superior performance in overall accuracy, average accuracy, and Kappa coefficient. HSIMamba's efficiency and performance make it a promising candidate for hyperspectral image analysis in remote sensing.