A neutronics benchmark experiment on vanadium, which is a low activation fusion reactor material, was conducted by using the D-T neutron source facility ofFNS/JAERI. Neutron spectra, dosimetry reaction rates, ,-ray spectra and ,-ray heating rates were measured in a vanadium experimental assembly. Benchmark tests for four evaluated nuclear data files were performed by analyzing the experiment. As a result, the following problems were pointed out in view of accuracy of fusion reactor designs. (1) JENDL-FF and JENDL-3.2: Total cross section should be reexamined especially at "'2keV. (2) ENDF/B-VI: Double differential cross sections for 14 Me V neutrons should be revised because of the isotropic angular distribution for continuum neutron emission. Gamma-ray production cross sections are too small and discrete ,-ray peaks are not represented clearly. (3) EFF-3: Gamma-ray production cross sections are too large.