<p>The world population mainly depends on the plants and the agricultural sector for food. Plant diseases are the main problem faced by the agricultural sector, as they affect the yield of crops. Among the plant diseases, the plant leaf diseases are the most common. These diseases show visible symptoms gradually over the leaves. Detection and classification of plant leaf diseases in the early stages is necessary to be taken to prevent the plants. The manual detection and classification of plant leaf diseases is a complicated task as it is time-consuming and contains errors. For effective plant leaf detection and classification, this paper proposes a Weighted Random Logistic Support vector Bayes based Hunter Prey Search (WRLSB-HPS) algorithm. The WRLSB-HPS method is trained on the plant village dataset for the validation of its performance. The input images are pre-processed using noise elimination, image resizing, normalization, image sharpening, and image blurring and then they are allowed into the detection and classification phase. Feature selection and classification are carried out using the four machine learning techniques Logistic Regression, Support Vector Machine, Naive Bayes, and Random Forest. The weights of these machine learning techniques are allowed into the weighted ensemble module to increase the accuracy of detection. The weights obtained from this ensemble WRLSB model are optimized by utilizing hunter-prey optimization along with an initial search strategy. The final classification output is generated with predicted class labels that denote the diseased or healthy plant leaf. The validation of this technique is done using performance measures like precision, accuracy, and F1-score. Comparison and evaluation results reveal that the WRLSB-HPS method achieved higher performance over existing methods with an accuracy of 98.4%. precision 98.2%, recall 97.9%, and 97.5% F1-score.</p>