Tomato plants are one of the most important component of global crops, but they are constantly affected by various kinds of diseases that has the potential to drastically damage the yield and quality. Timely diagnosis of these diseases is crucial in averting huge economic losses and ensuring the management of crop health. This paper conducts an elaborate review on disease detection strategies for tomato plants using various techniques, approaches, and algorithms. We used the InceptionV3 model for the already existing dataset instead of EfficientNet-B0, which obtained us better results. Our research examines several DL models in an effort to show how their performance can be improved in the classification of disease stages so that the disease factors are well addressed in a timely manner. The purpose of the proposed method is to enable accurate stage-wise disease detection of the tomato plant, using image-based diagnosis, as deep learning provides accurate automated plant disease detection and supports sustainability by reducing chemical use. Through comparative analysis, we establish the most suitable structures for DL models for disease accuracy detection, including one that was custom designed and yielded good results of 98.55% of test accuracy in EfficientNetB0 and 98.85% of test accuracy in InceptionV3 for disease detection.