The economic development and food security of Agro-based nations are significantly influenced by agricultural planning. Around the world, crop pests are responsible for enormous losses in the economy, society, and environment. Identification of pests is a huge challenge in agriculture, various pests require different pest control methods thus it has become crucial to accurately identify them. Sometimes, insect pests are misclassified leading to the improper use of pesticides, which is harmful to agricultural productivity and the environment. Because of their computational complexity and shortage of data, pest identification approaches have comparably poor accuracy in recognising and categorising pests. A lot of agricultural experts are intrigued by deep learning models, as they have demonstrated a lot of potential in image recognition. The main goal is to create a Convolution Neural Network (CNN) model that analyses and classifies the pests based on the image analysis of the pest. The best model is chosen by comparing all the accuracies and losses of several image classification models. The accuracies for the deep learning models along with the optimization techniques are calculated and the highest accuracy is seen for the EfficientNet-B0 model. The proposed model is evaluated using Kaggle standard dataset ‘Pest-Dataset’ to classify nine pests: aphids, armyworms, beetle, bollworm, grasshopper, mites, mosquito, sawfly, and stem borer. The proposed model may help researchers and agriculturalists in minimizing commercial and agricultural loss. Hence predicted pest classifies the pest class.