Abstract: Meteorological fields represent a multidimensional dynamic environment; therefore visualization is a means study inherent regularities in these phenomena. In this paper, aiming to visualize the worldwide, multi-dimensional, multi-scale, massive characteristics of meteorological field data, a voxel object-oriented data model and a voxel based multi-level indexing mechanism were designed. A visualization method for 3D meteorological fields is proposed, including an arrow model based visualization method for vector fields and improved 3D texture-mapping based volume rendering method for scalar fields. Visualization and spatio-temporal retrieval experiments on a Virtual Globe platform were carried out to demonstrate the feasibility and effectiveness of the proposed methods.