蒙特卡罗与热工水力的耦合计算是目前反应堆数值模拟的重要研究方向，在蒙特卡罗方法连续能量点截面的基础上结合热工程序的温度反馈，反应堆中子计算的准确性得到大幅提高。为了提高计算精度，堆芯模型分辨率也需进一步提高，相比于组件均匀化模型，pin-by-pin的建模方式能够获得更好的结果。利用蒙特卡罗程序JMCT与子通道程序COBRA-EN实现了蒙特卡罗-热工的内耦合，内耦合方式通过内存进行数据传递，其计算效率及安全性均优于外耦合方法。随后利用NURISP项目迷你堆的pin-by-pin模型对耦合程序进行验证。计算结果与同类耦合程序相似，验证了程序的准确性。同时，对耦合过程的收敛性问题进行了初步分析。

Monte Carlo is a statistical method widely used in solving particle transport problems. A Monte Carlo code has the advantages of much flexible geometry and high fidelity. Taking advantage of this, reactor core analysis can be solved with high fidelity, although much computing cost is needed. With the feedback of a thermal hydraulic code, the core in hot full power condition on steady state can be computed by a Monte Carlo code. In this paper, JMCT, a Monte Carlo code, and COBRA-EN, a subchannel code, are coupled through the method of internal coupling. Picard iteration is used between neutron transport calculation and thermal-hydraulic calculation. The HFP steady-state calculation in a mini-core benchmark in NURISP project is used to validate the coupling code. The result agrees with that of SERPENT2/SUBCHANFLOW and TRIPOLI/SUBCHANFLOW, verifying the accuracy of this code.