Modern landscape design not only needs creativity but also needs auxiliary tools that can predict the design effect, so as to ensure that the deficiencies can be found through the renderings before the landscape is completed, and targeted rectification can be carried out. At present, the research of landscape planning and design assisted by virtual reality technology in China is basically in its infancy. The rapid development of information-based computer technology, powerful 3D modeling, and solid rendering and animation functions have created a good environment for landscape design, so landscape design is inseparable from the help of computer-aided design technology. However, the conventional method to model and simulate the landscape is rather time-consuming. Based on the research on the application of computer-aided design technology in landscape design, the computer-aided design technology is briefly explained, and the application of computer-aided landscape design is explored step by step.

The layout and planning of urban landscape has a strong correlation with urban land utilization rate and ecological environment index. Urban landscape architects have a hard time dealing with these interrelated factors. This study uses a multicriteria constraint algorithm to optimize the relevant factors in urban landscape layout and planning. The convolutional long short-term memory (ConvLSTM) method was used to extract temporal features for urban landscape layout and planning tasks. Compared with the multicriteria algorithm without constraints, the multicriteria algorithm with constraints can better optimize the layout and planning tasks of urban landscape, and the maximum error of this method is only 1.96%. At the same time, the distribution of errors is more uniform under the multicriteria constraints, and it is all within 2%. The fusion of the multicriteria constraint algorithm and the ConvLSTM algorithm can better predict the relevant factors of the urban landscape layout, and the linear correlation coefficients of the three relevant factors have reached a high standard.