In order to improve the modeling accuracy of urban gardens and other types of landscapes, a 3D landscape modeling method is designed based on the laser 3D point cloud technology. Laser three-dimensional point cloud stereo non-contact measurement technology is used to obtain the three-dimensional coordinates of the landscape surface data points. The point cloud data of all angles are unified in a coordinate system, the longest distance between sequential points is used as the filter standard, and the point is set beyond the standard. In order to fix the endpoints and smoothly process the image point cloud, the triangular mesh parameterization strategy is used to map the three-dimensional mesh model to the two-dimensional plane to obtain the texture coordinates of the characteristic points. The harmonic mapping algorithm is used to solve the texture coordinates of the unconstrained points. By partially adjusting the strategy, the texture of the point cloud data is optimized and the final landscape model is obtained. The experiment collects three-dimensional data of a landscape in the study area, and combines the effects of the landscape model and the evaluation index to find out. The proposed method can effectively develop the model with high accuracy, and the detailed information of the model is relatively complete.