The consequences of extreme weather changes on crops, the growing global population, and numerous illnesses that emerge at a severe level all have an impact on agricultural production and food security. Farmers employ expensive disease management strategies that sometimes outweigh output losses. Because it would be too expensive for them to install crop protection measures at the early beginning of the illnesses, which results in very little advantages, the farmers would only act when it was too late to stop the disease from spreading. It was suggested that computer orders using technology for calculating vision may be beneficial in a range of original-globe situations. It may be used in crop security strategies since it employs artificial intelligence, graphics, and image processing to detect illnesses, pests, and malnutrition. Machine learning (ML) approaches are being utilised in the ongoing study of disease detection from plant photographs. For recognising paddy leaf disease, a unique deep neural network (DNN) classification model is employed using plant picture data. Existing system like CNN, TL and ANN gives the mere accuracy results. There is drawback in identifying the accurate crop diseases. The proposed hybrid ECNN and CSA algorithm give more accuracy when compared to the existing system. It increases the accuracy level by 95%. The hybrid ECNN & CSA based designs shown amazing success in crop disease prediction and categorization using images. These models require a lot of training data, though, and they are computationally costly. The proposed enhanced hybrid deep learning based hybrid algorithm model identifying the diseases in the paddy leaves in a most significant manner.