Cassava stands as a pivotal food crop in Sub-Saharan Africa. It is susceptible to several diseases such as bacterial blight, brown spot, green mite, and mosaic disease. There is need to identify and detect these diseases. Deep learning algorithms, particularly Convolutional Neural Networks (CNNs) are employed by numerous researchers for disease detection. However, overfitting is a major factor that hinders the overall performance of Convolutional Neural Networks (CNNs). Therefore, in this study, an optimized Convolutional Neural Network (CNN) using Artificial Bee Colony (ABC) algorithm was developed to enhance the accurate and timely detection of some cassava diseases. The model was developed by employing Artificial Bee Colony (ABC) algorithm to optimize the parameters of Convolutional Neural Network (CNN). The dataset used was downloaded from data.medeley.com, comprising healthy and infected cassava leaf images of four distinct diseases (bacterial blight, brown spot, green mite, and mosaic disease). It was split into training and testing set of 70% and 30% respectively. The images were pre-processed, Principal Component Analysis (PCA) was employed to extract relevant features and implemented using python 3.11.4. The model was evaluated in terms of Precision, Sensitivity, Specificity, False Positive Rate (FPR), Accuracy, and Average Recognition Time (ART). The CNN model recorded 85%, 85%, 85%, 4.5%, 84%, 0.161007sec for Precision, Sensitivity, Specificity, FPR, Accuracy, and ART respectively while the ABC-CNN model recorded 95%, 89%, 97%, 2.5%, 90%, 0.157266sec for Precision, Sensitivity, Specificity, FPR, Accuracy, and ART respectively, underscoring the performance of the models. Based on the results, the developed ABC-CNN model performed better across all the performance metrics compare to the CNN model. The developed ABC-CNN model is recommended for the agricultural field to ensure accurate and timely detection of cassava diseases.