Understanding fluid flow and the fluid retention capacity of a partially saturated porous medium needs an accurate quantification of capillary pressure. In a porous medium like soil, multiple components can coexist, like solid, non-wetting fluid, and wetting fluid. The multicomponent Lattice Boltzmann Method (LBM) has emerged as a powerful tool for simulating fluid flow in porous media, complementary to micro-CT characterization of real samples. Despite the method’s capability to handle complex geometries, the accuracy of capillary pressure calculation in multicomponent LBM simulations (e.g., via Shan-Chen pseudopotential) is largely influenced by a number of factors. These include insufficient or excessive filtering of interfacial lattices, inadequate resolution, and incorrect simulation parameters. In this paper, we propose a method based on the component’s number density at a lattice site, for improving capillary pressure calculation. To assess the correction method, a range of simulations were conducted, from a single droplet in a duct to different unit cells of regular, idealized granular systems and porous media with complex pore geometries. By adjusting the density thresholds for differentiating fluid bulk from interfacial lattices, it is ensured that lattices in the interfacial zone do not contribute to the calculation of capillary pressure. The measurements are fine-tuned to determine the value at which the capillary pressure reaches a plateau, indicating the appropriate threshold for a given multiphase configuration. A main contribution is a grid resolution convergence test, aimed at validating the method's accuracy by carefully scaling parameters to maintain the same fluid properties. The proposed method can eliminate the errors that arise from not filtering out interfacial lattices in the capillary pressure averaging. Overall, this study offers a practical solution for improving the accuracy of capillary pressure calculations in unsaturated soil, as well as valuable insights for researchers working with the multicomponent Shan-Chen LBM in complex geometries.