Abstract: Pomegranate may be a broadly developed planting within Indian. This Profoundly useful natural product is contaminated by numerous bugs and illnesses which cause extraordinary temperate misfortunes. Distinctive Shapes of the pathogens maladies within leaves, Stems and of natural products are show. A few of the maladies that influence pomegranate natural products are anthracnose, heart decay and bacterial curse. There’s a require for malady control procedures to incorporate opportune activity on the created infections. Hence, there’s a require for shrewdly and self-learning acknowledgment frameworks to identify these illnesses within given Time. This think about is pointing to classifying of pomegranate natural products in twice classes ordinary and unusual utilizing CNN&LSTM procedure. This investigate work utilizes with cross breed CNN&LSTM procedure to distinguish four sort of maladies show within the pomegranate natural products and classifying all of them within 4 different classes. The comes about gotten utilizing CNN&LSTM are at that point optimized utilizing dragonfly calculation. The highlights like color, surface and shape of the natural products are gathered & bolstered into the half breed CNN&LSTM. The datasets with the classifier is understood like an exceed expectations record which is at first pre-processes utilizing outline diminish procedure and dimensionality carried utilizing foremost component investigation and Discriminant investigation.