Fission reaction rate is an important index for validating and checking the neutron transportation and fission power in nuclear engineering. The experimental data can be used in benchmark validation of cross sections, and in studying the correlation of fission power with the thickness of uranium sphere shell. There are five assemblies of depleted uranium shells used in this work, the inner radii of which are all fixed at 13.1 cm, while their outer radii are 18.1, 19.4, 23.35, 25.4 and 28.5 cm, respectively. The D-T neutron source is generated in the center of the assemblies, the yield of which is about 3 × 10 ¹⁰−4 × 10 ¹⁰s ^–1. In horizontal plane across the center of the assemblies, the fission rates at positions along the radial direction are measured in the direction with 45° inclining with respect to the incident D ⁺beam. Due to the disturbance to assemblies and neutron field, the activation foil of uranium is a suitable choice rather than fission chamber or capture detector. The material of activation foil is the same as that in the experimental assemblies. Considering the accurate fission yield of ¹⁴³Ce, the objective nuclides are selected. The total fission yield of ¹⁴³Ce is contributed by ²³⁸U and a little ²³⁵U. For calculating the total fission yiled of ¹⁴³Ce, the neutron energy range of 0−15 MeV is divided into eight subranges. By measuring the 293 keV gamma rays from the fission product ¹⁴³Ce in activation foils with a TRANS-SPEC-DX100 HPGe detector, with a relative efficiency 40%, the fission rates and the trends at positions along the radial direction in the five assemblies are obtained based on the ¹⁴³Ce fission product yield. The fission rate ranges from 5.28 × 10 ^–29to 7.58 × 10 ^–28sn ^-1·nuclide ^–1, with the relative uncertainty in a range from 6% to 11%. The Monte Carlo transport code MCNP5 and continuous energy cross section library ENDF/BV.8 are used for analyzing the fission rate distribution in the assemblies, and the experiemtal configuration, including the wall of the experimental hall is described in detail in the model. The calculated results are compared with the experimental ones and their agreement is found to be in an uncertainty range.