This paper presents a computer-aided cloud-analysis approach by effectively modeling the integration of heterogeneous satellite-observed data and remote sensing images. First, automatic cloud detection and tracking methods are proposed to identify the georeferenced cloud objects in satellite remote sensing images. Then, a data integration modeling mechanism is designed to collect meaningful properties of those detected clouds by integrating the heterogeneous satellite-observed data and imaging into a unified cloud database. Finally, based on the integrated global data schema, a two-phase data mining method employing the decision tree algorithm is implemented to analyze and forecast the meteorological activities of all the cloud objects. Experimental results have shown that the proposed data integration model can effectively extract and synthesize all the useful information from heterogeneous data sources to generate a unified view of knowledge, on the basis of which the evolvement trends of clouds can be analyzed properly.