The development of agriculture is related to the stability of the country’s economy. Changes in agricultural productivity have been one of the more complex current agricultural problems. To improve agricultural productivity, this study built a pest detection model based on agricultural information technology. This model used agricultural information technology to analyze agricultural pests and diseases. The results showed that crop yields were higher when using the research model in crop productivity. The training and testing of three crop pest and disease data showed that, in different models, the accuracy of the proposed model was 1.22%, 2.47%, 3.47%, and 1.68% higher than that of ResNet50, ResNet101, MobileNetV1, and MobileNetV2, respectively. In addition, the proposed model could process the data stably and reduce the error value. The utilization of an enhanced convolutional neural network method demonstrated the enhancement of green vegetables yield, while concomitantly reducing the impact of pests and diseases and improving the efficiency of agricultural production. This research provided a good reference value for the future improvement of agricultural production efficiency. © (2024), (Bio Tech System). All rights reserved.