A three-dimensional model processing method, an electronic device, and a storage medium are provided, which are related to fields of deep learning, augmented reality, and the like. The specific implementation includes: generating a target virtual three-dimensional model based on a target model template; generating a target picture based on graphic and textual information and a preset network; and determining a superposition processing result of the target virtual three-dimensional model based on the target virtual three-dimensional model and the target picture.