In today’s fast-paced world, vehicle classification is an important task due to its wide range of applications, such as Intelligent Transportation Systems (ITS), autonomous driving and traffic management. This study proposes a comparative analysis on the different variations of CNN; Custom CNN and pre-trained models like ResNet-50 and MobileNetV2. The main objective of the study is to find the role of data augmentation, whether it contributes effectively towards the models after training. Data augmentation is a method to increase the size of data using different methods and the models are trained twice on original data and with data augmentation to understand the gap in their performance. It illustrates that the risk of overfitting is reduced and the ability of the models to generalize has improved with data augmentation. The performance metrics which is used to evaluate the models are accuracy and loss curves of the model. The pre-trained model MobileNetV2 has acquired a remarkable accuracy of 99.52%, which is the best-performing model. This study provides a transparent approach in selecting the most efficient model for vehicle classification.