From both fundamental and applied points of view, fragment mass distributions of fission are important observables. We apply the Bayesian neural network (BNN) approach to learn existing neutron induced fission yields and predict
unknowns with uncertainty quantification. By comparing the predicted results and the experimental data, the BNN evaluation results
are quite satisfactory on distribution positions and energy dependencies of fission yields. Predictions have been made for the fragment mass distributions of some actinides,
presuming that this might help for future experiments.