This research addresses the computational complexity of data fusion under the Dempster-Shafer mathematical theory of evidence as well as the Dezert-Smarandache theory. As earlier research has shown, the use of these theories results in better data fusion results. However utilization of these theories is hampered by computational complexity when dealing with large scale problems. This paper presents a new approximation algorithm allowing target classification and identification for a larger number of target classes with a reasonable execution time. The paper also presents the results of application of the developed method to classification of naval ships.