Hyperspectral images are very different from normal RGB images as RGB images constitute only three energy bands whereas a hyperspectral image contains hundreds of energy bands, therefore hyperspectral image classification predictive models are very complex and require specific parameter tuning for efficient working of the model. This paper proposes a hyperspectral classification framework that uses optimized parameters obtained via genetic algorithm. Different feature extraction and selection techniques like PCA is also used to reduce the excess dimensions of images. The proposed framework is implemented on the various machines and deep learning models to show the effectiveness of the proposed work.