Gooseberry farming is crucial to agriculture; the entire condition of these gooseberry plants is critical to ensuring optimal yield and fruit quality. The purpose of this project is to categorize gooseberry leaves to recognize various diseases utilizing an amalgamation of neural networks (CNN) convolution or Random Forest techniques. The powdery mildew, anthracnose, leaf spot, rust, leaf curl, aphids disease, septoria, leaf spot, along canker are the eight disease types being investigated. The classification method is thoroughly evaluated using precision, recall, F1-score, assistance, or accuracy criteria for each sickness class. The model performs well, with accuracy ratings ranging from 93.67 to 94.59%, recall percentages extending from 93.82 to 94.70%, and F1-scores ranging from 93.94 to 94.45%. The support values, which vary from 865 to 945, indicate how instances are distributed between classes. Furthermore, the model’s general precision is remarkable, at 98%. Macro, weighted, and microscopic averages complete the assessment by aggregating measurements from all classes. The macro-average shows balanced achievement in precision, recall, and overall F1-score, all of which average 94.16%. The weighted average adjusts for class-specific support, yielding consistent results of 94.17% across all variables. The micro-average provides an overall assessment of the algorithm’s efficacy of 94.17%, which is consistent with both the macro and weighted averages. This study advances precision agriculture by proposing a dependable approach for the early detection and categorization of individual gooseberry leaves.