Abstract. The analysis of k eff multiplicative factor of 100%-MOx cores and plutonium solutions showed a systematic discrepancy between calculation and experiment. The aim of this paper is to propose an improved version of the JEFF-3.1 239 Pu evaluation in the low-energy neutron range. We propose here slight modifications of the evaluation file, consistent with the uncertainties of differential data measurements in sub-thermal/thermal and epithermal ranges, namely: modifications of the sub-thermal fission, capture and total cross sections, and a revised evaluation of ν p from sub-thermal up to 20 eV using a phenomenological formalism. The proposed modifications are described and tested against integral measurements.