Pests pose a significant threat to cotton yields, leading to economic losses for farmers. This paper conducted a comprehensive literature review, examining the various types of pests affecting cotton crops both globally and specifically in India. We analyzed the actual yield losses ascribed to these pests as well. Furthermore, we reviewed a variety of literature on pesticide recommendations and cotton disease detection. In this paper we have chosen four Deep Learning (DL) models—VGG19, ResNet50, ResNet152V2, and InceptionV3 selection of this model was based on literature review. These models were trained and tested on a dataset of 2,293 images comprising affected and non-affected cotton plants and leaves. The output of our evaluation demonstrates how useful these models are in detecting diseases in cotton plants. Ultimately, the AI-based automated recommendation system allows farmers to use this information to make informed decisions about using pesticides resulting in better crop health and higher yield of cotton.