ABSTRACT This paper aims to enhance the effectiveness of moral education teaching through an innovative approach. By integrating neural network technology, a data-model-driven three-dimensional initial model construction algorithm was proposed to identify the characteristics of moral education classroom teaching. The 3D space base point and initial value of shape parameters were obtained based on color point cloud data, and the 3D space coordinates of all the vertices of the model were calculated using the Euclidean distance algorithm, resulting in the construction of an initial 3D model. Furthermore, a model of Innovative Moral Education (IME) in higher education was developed based on neural network, and a framework analysis was conducted. Results from simulation teaching research indicate that the IME teaching mode based on the proposed neural network model has a significant practical effect. Overall, this study offers insights into an innovative approach to moral education teaching and contributes to the development of effective pedagogical practices.