Summary The purpose of this study was to develop a mathematical model which would be easy to use and which could provide an objective classification of experimental brain tumors. 250 brain tumors (glioblastomas, astrocytomas, astroblastomas, oligodendrogliomas, ependymomas, fibrosarcomas and giant cell sarcomas), induced in mice, were studied. In every case the same 50 histological characteristics were analyzed. For astroblastomas, oligodendrogliomas and ependymomas the diagnosis given by the neuropathologist and that given by the computer corresponded in 100 % of the cases. However, some fibrosarcomas and giant cell sarcomas were not grouped in the same diagnostic category by the computer as that assigned by the neuropathologist. Thus an overall efficiency of 94.35 % was obtained. This simple and efficient system of classification allows for further statistical analyses and makes it possible to correlate the results presented by different authors, thus providing for unity in all those works concerned with experimental neurooncology.