The stiffness ratio between an inclusion and the surrounding tissue provides critical information for tumor classification. Malignant tumors are usually harder than benign ones. Accuracy and efficiency of computing tissue stiffness depends on how external excitations are applied to the tissue and what kind of biomechanical model is used. In this paper, a lateral exploration strategy combined with an inverse biomechanical model based on an artificial neural network has been proposed to identify inclusion properties. The experimental results showed that the proposed method was able to predict the inclusion properties with better accuracy and significantly improved computational efficiency as compared to the conventional indentation method.