In the era of big data information digitization, digital terrain editing technology is an important research content in the virtual geographic environment, 3D landscape simulation system, which has significant value in digital planning and design practice of cities and villages, regional road networks, and landscape gardens. This paper focuses on deep learning-based digital terrain interactive editing technology for landscape planning and design. Firstly, it introduces the concept and principle of digital terrain interactive editing technology, constructs the model and terrain interactive editing algorithm, and then analyzes three systems’ digital geography interactive data from two aspects of accuracy and convergence. The results show that A system planning digital landscape garden landscape accuracy, planning accuracy, and convergence are small, and the planning accuracy reaches a minimum of 88% and a maximum of 95%. The intersection reaches a minimum of 75% and a maximum of 87%. Compared with the other two systems are high, so it can be seen that the digital interactive terrain editing for landscape planning of this design has high planning accuracy and convergence. Digital interactive editing technology is developing rapidly in China’s landscape architecture discipline research and professional practice applications. It has shown excellent development potential and application value in the landscape planning and design industry. This study improves the accuracy and efficiency of digital interactive editing technology and has great historical significance for the development and progress of the Chinese architectural design industry.