Based on a dynamic one-dimensional model, lattice Boltzmann method (LBM) was applied to simulate cafestol and kahweol extraction from green coffee beans in high-pressure system. The model entailed coupled partial differential equations to extract concentrations in fluid and solid phases, having intra-particle diffusivity, solid-fluid partition coefficient and fluid-phase diffusivity as model parameters. LBM was implemented in D1Q2 lattice by assigning particle distribution functions to concentrations in both phases. Equilibrium distribution functions were defined considering stationary solid phase and diffusive convective transport in fluid phase. By combining the gradient descent algorithm to the numerical procedure, model parameters were best-fitted against experimental extraction data at 60 degrees C, 80 degrees C and 100 degrees C. As temperature increased, fluid-phase diffusivity decreased whereas intraparticle diffusivity and partition coefficient increased. Though neglected in convective-dominant models, sensitivity analyses suggested that fluid-phase diffusivity is more influential than other parameters. Expected behavior of extract yield curves was reproduced in all LBM simulations. (C) 2015 Elsevier Ltd. All rights reserved.