Abstract: The cloverleaf is an important component of urban traffic planning, at the same time it is one of the most complex features in the city. The 3D visualization problem of cloverleaf is a serious topic in cyber-city construction. To the reason of its complexity in construction and topology, results made by many traditional unified modeling functions are not able to reflect its spatial significance, while the abstract expressions of some commercial software cannot provide intuitional investigation and interactive operation. To solve these problems, this paper presents a method to build 3D visualization model of city cloverleaf based on airborne laser scanner data. The prime features include: 1) the method enters the modeling step after the segmentation of cloverleaf datasets. We provides a automatic segmentation algorithm which divide the complex structure of road divaricates and cross into sections of simple construction. This operation laid a solid foundation for the next work; 2) we use all-constraint triangulation algorithm to build the model of cloverleaf, the constraint section is chosen from the contour; 3) we can detect and reconstruct the absent part of the data occluded by the upper structure, it is based on the model features of sections combines with priori knowledge. The results shows our function which build the 3D model after the segmentation process can worked out satisfactorily, the all-constraint triangulation algorithm keep well of the spatial characters of cloverleaf, more importantly, it can help to rebuild the part occluded by the upper structure. The analysis further explains the feasibility and efficiency of our method.