Anticipating the type of orthodontic treatment needed holds significant weight in patients' decision-making processes. In recent times, several deep learning techniques have shown impressive results in several computer vision tasks, especially in image segmentation, image classification, image identification, etc., using convolutional neural networks. This study aimed to develop a model using convolutional neural networks for predicting orthodontic malocclusion types, which are very difficult to work on using traditional image processing techniques. Currently, proposals to use orthodontic software in dental clinics have been put forward; however, they lack the ability to comprehensively analyze complex patient data, including genetic factors, craniofacial features, and treatment outcomes, potentially leading to sub-optimal treatment decisions. Therefore, the use of machine learning in orthodontics remains largely unexplored due to it being a very vast field. In this study, dental images of three thousand six hundred (3600) patients were used to predict the type of malocclusion. These included images both before and after the treatment. The results of this study demonstrate the effectiveness of convolutional neural networks in accurately classifying different types of orthodontic treatment. The findings reveal the potential for machine learning to assist orthodontists in treatment planning decisions, providing valuable decision support to orthodontists and improving patient outcomes, thus opening avenues for further advancements in orthodontic care.