Tomato is an essential crop, grown across many regions and consumed globally. However, tomato plants are susceptible to diseases, which can significantly affect the yields of crops leading to substantial losses. Early detection of these diseases is crucial to prevent losses. Previously, conventional machine learning methods used to identify plant diseases proved notably ineffective. However, with the introduction of deep learning, specifically convolutional neural networks, detection results have shown remarkable enhancement and precision. In our work, we conducted multiple experiments using images captured in controlled and natural environments, sourced from two different sources. Our proposed approach achieved the best accuracy of 98.50%.