Abstract: The core inventory calculation of a pebble-bed high-temperature gas-cooled reactor aims to capture the instantaneous changes and spatial distribution of nuclides. This provides a foundation for nuclear safety aspects such as radiation protection design, radioactive waste management, and environmental impact assessments of nuclear power plants. It also serves as a prerequisite for analyzing fission product source terms. The multi-pass fuel management and the random movement of fuel pebbles in pebble-bed high-temperature gas-cooled reactor pose significant challenges for core inventory calculations. Traditional methods based on lumped parameters fail to account for the spatial distribution in radioactivity source term and cannot capture differences in core inventory accumulation among fuel pebbles. In this paper, a refined calculation method for core inventory calculation at the level of individual fuel pebbles was proposed by leveraging core power distribution and random sampling techniques. By randomly sampling the position and power history of individual fuel pebbles and performing burnup calculations, the variation in nuclides inventory was obtained. These inventory state points were distributed throughout the core based on the pebbles’ positional histories. The state points generated by multiple pebbles represented the total core inventory. Using HTR-PM (high-temperature gas-cooled reactor pebble-bed module) as a case study, the differences between the refined and traditional methods was compared, and the impact of pebble flow velocity distribution on core burnup distribution was examined. The results show that the refined method yields a slightly lower total core inventory of fission products compared to previous methods, with a greater average burnup in the reactor core. Previous methods are more conservative, the refined method incorporates more factors influencing core inventory calculations, leading to more precise results. Compared to the uniform flow velocity model, the non-uniform flow velocity model, due to smaller differences in average burnup increments among various core channels, results in a more uniform axial burnup gradient. This reduces the differences in burnup among fuel pebbles at discharge. For future research, coupling the refined method with pebble flow simulation programs could produce a more realistic velocity distribution, enabling more accurate core burnup and discharge burnup predictions. Such advancements hold significant implications for radioactive waste management and radiation shielding design.