This thesis proposes a novel deep learning architecture called ValidNet to automatically validate the correctness of a 3D surface registration outcome. The performance of many tasks such as object detection mainly depends on the applied registration algorithms, which themselves are susceptible to local minima. Revealing this tendency and verifying the success of registration algorithms is a difficult task. We treat this as a classification problem, and propose a two-class classifier to distinguish clearly between true positive and false positive registration outcomes. Our proposed ValidNet system deploys a shared multilayer perceptron architecture which works on the raw and unordered point cloud data of scene and model points. This network is able to perform the two fundamental tasks of feature extraction and similarity matching using the powerful capability of a deep neural network. Experiments on a large synthetic dataset show that the proposed method can effectively be used in automatic validation of 3D surface registration.