Leaf diseases are a wide range of conditions affecting plant leaves, frequently caused by fungi, bacteria, and viruses. The leaf diseases may have an impact by reducing agricultural yields, production, and quality. One of the main crops produced on a vast scale is the tomato. This study employs deep learning techniques and computer vision technology to help anticipate crop diseases to support agriculture. The proposed algorithm consists of depth-wise separable layers followed by convolution, max pooling, and deconvolution layers to improve the classification of diseases. Experiments are conducted on the well-known Plant Village dataset to examine the effectiveness of the suggested approach. The layers of the proposed model are more computationally efficient. Using fewer parameters than the traditional and existing models, the recommended model achieves an overall accuracy of 87.4% in classifying and identifying tomato leaf diseases into 10 separate groups.