Benchmarking criticality experimental data can support the verification of theoretical design methods, design procedures and nuclear data for new reactors. In order to fill the gap of criticality experimental benchmark data for small lead-cooled reactor in China, criticality experiments were carried out on the YBS Criticality Experiment Facility lead reactor core in China Institute of Atomic Energy(CIAE), By modifying the loading method of fuel and moderator in the core components, the fast neutron energy spectrum and the Mixed neutron energy spectrum were realized, And several sample criticality experiments were carried out to test the material properties of various experimental samples under the two energy spectrums. Based on the GUM uncertainty analysis process, the integrated uncertainty of the criticality experimental data was obtained by considering all possible sources of experimental uncertainty, the detailed benchmark models were final built and applied in the applicability validation of four combinations of Monte Carlo software and nuclear data libraries for keff calculation in the specific systems.