This paper presents the implementation of vehicles classification using Extreme Gradient Boost (XG Boost) Algorithm to improve the accuracy in the vehicle classification with respect to the shapes and features. Generally, Classification based on different parameter such as hefty, classes, structures, extracting features, segmenting the images and semantic classification are being challenge to incorporate in Machine Learning. In order to overcome this barrier, XG boost algorithm has been implemented to achieve the high performance vehicle classification from the large scale surveillance dataset. The experimental results shows that the accuracy is improved in the vehicle classification with standard resolution image.