Classifying crops and leaves into healthy and unhealthy categories is central to modern agriculture as it is key for early disease diagnosis which helps improve the yields. This work puts forward an AI-based technology for classifying and measuring the healthy and unhealthy bell pepper сrops and leaves through Mobile Net V2 techniques. The Mobile Net V2 model was developed on a collection of 6086 training and validation samples with 1519 samples and was shown to possess high classification accuracy. The experimental results indicate the model achieved an overall accuracy of 98.2% in the training dataset and 98% in the validation dataset. The model also performed well in different categories: Crop Healthy, Crop Rotten, Leaf Healthy, and Leaf Unhealthy with precision, recall, and F1-scores higher than 0.90. For example, the validation set had a macro-average F1 score of 0.95 which points to efficient generalization. These results confirm the AI-based approaches in automating crop health detection and disease classification models are efficient. The results of the research claim that AI-based classification models can be an aide that solves real-time agricultural problems for farmers, which helps them assess crop health. Further research will be aimed at real-time application, IoT integrations, and extending the range of crops monitored for efficient agricultural disease detection and management.