Quantum computing in machine learning gives us a new tool to address few complex problems in the domain of artificial intelligence (AI). In this work, we use the hybrid, i.e., a combination of quantum and classical Quantum convolution neural network (HQCNN) in a hierarchical way for image classification. A four class classification is obtained hierarchically using three HQCNNs. In phase 1, we perform binary classification using HQCNN1 to get the two class classification with each class a mix of two individual classes. We then further classify each of the images into their individual categories with the use of HQCNN2 and HQCNN3 in phase 2 by performing two more binary classifications. Two experiments are conducted to show the efficacy of our proposed approach. In the first experiment we use X-ray images of organs consisting of Chest, Hand, Ankle, and Knee images which are taken from UNIFESP dataset. In our second experiment we use Deer, Horse, Airplane, and Ship images drawn from CIFAR10. We implement our proposed approach on Pennylane Software Development Kit (SDK), a simulation tool that can be implemented on a classical machine. Comparison of results with the classical method clearly indicates that the training using the proposed approach takes fewer number of epochs with improvement in accuracy and the convergence is also stable when compared to the classical approach.