Abstract: The design and safe operation of nuclear reactor are inseparable from high-precision and highresolution simulation program. With the development of calculation methods and the progress of computer hardware, highfidelity numerical reactor technology becomes a research hotspot. However, due to the complexity of the objective world and the limitation of human cognitive level, there is uncertainty in any measurement and mathematicalphysical modeling process. Therefore, the simulation results need to provide the corresponding uncertainty as the confidence interval to meet the requirements of high confidence in numerical reactor technology. In order to promote the development of sensitivity and uncertainty analysis of reactor system, OECD/NEA (Organization for Economic Cooperation and Development/Nuclear Energy Agency) issued the UAMLWRs (uncertainty analysis in modelling of light water reactors) benchmark. The exercises are carried out in 3 phases that is neutronics phase, core phase and system phase. In the neutronics phase, the uncertainty of nuclear data is transferred from the evaluation nuclear database files (ENDF) to multigroup micro sections, and then to macro fewgroup sections and key parameters of steadystate core calculation. The core phase includes three independent physical processes which is fuel performance analysis, timedependent neutronics and bundle thermalhydraulic. The system phase includes multiphysics coupling calculation, and finally gives the best estimation plus uncertainty (BEPU) result. In order to provide confidence interval while giving numerical simulation results, the uncertainty quantification of fuel performance, assembly burnup and thermalhydraulic analysis was carried out based on the second phase (core phase) of UAMLWRs benchmark. Using the uncertainty analysis platform NECPUNICORN independently developed by Xi’an Jiaotong University, the light water reactor fuel analysis code FEMAXI, the lattice physics program NECPBambooLattice and the subchannel thermalhydraulic code CTF were coupled respectively. Firstly, according to the characteristics of different physical processes, the sources of uncertainty to be considered were analyzed. Then, according to the characteristics of sparse and unfilled rank covariance matrix of nuclear data, COST method was applied to reduce the sample size. It is found that for fuel performance analysis, boundary conditions, geometric parameters and material properties have significant effects on the uncertainty of fuel center temperature. For the burnup process, the nuclear data and manufacturing parameters have a significant impact on the uncertainty of eigenvalues, power distribution, twogroup constants and nucleon density. For thermal hydraulic analysis process, boundary conditions, geometric parameters and model coefficients have a great influence on the uncertainty of coolant and cladding temperature. For each physical field, quantifying the uncertainty of its input and output parameters is of great significance for quantifying the uncertainty of multi-physical coupling process of complex system.