Given the world's growing demand for energy, a combination of geological CO2 sequestration and enhanced oil recovery (EOR) technologies is currently regarded as a promising solution, as it would provide a means of reducing carbon emissions into the atmosphere while also leading to the economic benefit of simultaneously recovering oil. The optimization of injection strategies to maximize CO2 storage and increase the oil recovery factors requires complicated pore‐scale flow information within a reservoir system consisting of coexisting oil, water, and CO2 phases. In this study, an immiscible three‐phase lattice‐Boltzmann (LB) model was developed to investigate the complicated flow state with interaction between water, oil, and CO2 systems in porous media. The two main mechanisms of oil remobilization, namely, double‐drainage and film flow, can be captured by our model. The estimation of three‐phase relative permeability is proposed using the digital rock physics (DRP) simulations. The results indicate that the relative permeability of CO2 as calculated using our steady state method is not sensitive to the initial oil fraction if the oil distribution is originally uniform. Baker's (1988) empirical model was tested and found to be able to provide a good prediction of the three‐phase relative permeability data. Our numerical method provides a new tool for accurately predicting three‐phase relative permeability data directly based on micro‐CT rock images. Key Points A novel method for estimating of three‐phase relative permeability data is developed This method has the ability to evaluate saturation‐history effect Baker's model agreed well with the calculated relative permeability results