This paper constructs a three-dimensional object visual perception space using digital fusion technology to enhance the impact of art design works. A large amount of homogeneous or different visual art design information collected is fused, and data processing and fusion methods are used to reduce redundant data, establish a reasonable knowledge base to explain the information and derive the maximum support degree. The mean square error is calculated according to the adaptive weighted data fusion, and the trust degree function and likelihood function are introduced to obtain the total trust degree of visual art design innovation types. The results show that the beauty indexes become nonlinear, the fitted correlation indices are all greater than 70%, and the digital fusion technology can prompt the artworks to be more outstanding in visual effects.