Abstract: Numerical reactor is aimed to show the physical phenomena by means of high-fidelity numerical simulations. Monte Carlo method is commonly used in numerical reactor. JMCT (joint Monte Carlo transport) is a large-scale, high-fidelity, three-dimensional neutron-photon-electron transport Monte Carlo software system which is developed by CAEP-SCNS. For core physics calculation, JMCT has developed a number of functions. JMCT can make use of the multiple-temperature continuous energy data library which named Nuda based on ENDF/B-Ⅶ, and uses on-the-fly (OTF) Doppler broadening to improve the accuracy of multi-temperature cross sections. JMCT also has the ability of critical boron searching based on neutron balance, which is more efficient than the traditional k_eff-based linear interpolation algorithm. Based on the above functions, JMCT can calculate the reactor core of zero power. For hot full power cases, thermal hydraulic feedback is considered by coupling with a sub-channel code JTH. Furthermore, JMCT is coupled with a burn-up module named JBURN. Therefore, JMCT has the feather of transport thermal hydraulic and burnup coupling. VERA benchmark is a core physics benchmark proposed by CASL. It takes Watts Bar reactor as the original model and provides reference solutions for some problems with KENO-Ⅵ. It totally contains 10 problems, problem 1 to problem 5 are zero power problems, problem 6 and 7 are hot full power problems, problem 8 and 9 are depletion problems and problem 10 is a refueling problem. Problem 1 to problem 9 were modeled and simulated by JMCT, except problem 8. JMCT keff results are in good agreement with the reference values: The difference of keff is generally less than 20 pcm, and the max value is less than 150 pcm. The difference of control rod value between JMCT and the measured value is less than 60 pcm, and the results of boron coefficient and temperature coefficient are very close to the reference results calculated by KENO-Ⅵ. The radial power distribution of each assembly shows good agreement with KENO-Ⅵ and MC21 and the max difference is less than 1% in HZP cases. The above results verify the feather of criticality simulation of JMCT. In HFP cases, max difference of the radial power distribution of each assembly with MC21 is 2.27% and the critical boron concentration is similar to that of MC21. For depletion problem, the critical boron concentration difference in each step is less than 25 ppm. This work verifies the ability and accuracy of core physics simulation of JMCT.