Agricultural sector is seriously threatened by the transfer of disease from unhealthy to healthy plants which might infect whole fields if early detection is not achieved. Plant diseases reduce the yield losses that directly affect the food production systems resulting in economic losses. Disease detection techniques aid in the early identification of afflicted plants and efficiently expand disease identification to a range of plants in cost-effective manner. Farmers manually examine leaves to detect parasite, nutritional imbalances, or environmental conditions, but these methods are time-intense, error-prone, and rely primarily on pathologists. So, automatic detection models are essential for efficient plant disease prediction and farm disease management. Recently, Deep Learning (DL) models provide the efficient results on predicting all varieties of plant disease because of their capacity to properly anticipate and increase crop yields’ output in farms. This method may help farmers and plant pathologists to identify the infections earlier, lowering the time, cost, and labor necessary for real-time quality control. This research provides a complete overview of several DL techniques for detecting and classifying tomato infections using images. Initially, different tomato plant disease categorization systems developed by several researchers using DL algorithms are briefly examined. Then, comparison research is carried out to better comprehend the limits of current algorithms, to provide a novel technique for classifying tomato plant disease types, and to increase the agricultural productivity.