The agricultural market holds significant benefits to meet food demands by providing healthy and nutritious product. However, detecting crop insects remains a tough task for farmers, as pests cause considerable damage and degrade crop quality. Traditional methods of insect identification rely on experienced scientists to accurately recognize spiders depending on their shape structure and features, which can be limiting. Experiments were conducted on the Wang and Xie dataset, classifying nine and 24 insect categories using structural attribute and various artificial intelligence algorithms, including artificial neural networks (ANN), support vector machines (SVM), k-nearest neighbors (KNN), Persistence Model (PM), and convolutional neural networks (CNN). To enhance model performance, ninefold cross-validation was implemented. The CNN framework attained the greatest accuracy rate of 90.5% and 89% respectively with respect to Nine and twenty-four insect’s classes. The proposed framework demonstrated improved detection performance with reduced computation time on the aforementioned datasets. Compared to literature of the classification framework, the proposed method showed significant improvements in classification accuracy and computational efficiency, making it more effective for identifying insects in field crops. The classification accuracy results aid in early detection of crop insects, thereby reducing time and improving agricultural productivity and quality.