The worldwide combustion of fossils produces anthropogenic CO 2 , which has deleterious effects on global climate patterns. An ideal solution is to develop high-performance CO 2 capture and utilization technologies that can convert CO 2 into commercial products by means of electrocatalytic reduction with renewable energy. The design of electrocatalysts can be facilitated by accurate computational simulations based on density functional theory (DFT). Herein, commonly used models, from the computational hydrogen electrode model with constant charge assumption, to the explicit and implicit solvation model under constant potential condition, are reviewed. Thereafter, the applications of recent data-driven methods, such as neutral network and machine learning, are introduced. Also, based on DFT calculations, four composition based classes of electrodes are further discussed, including (1) bulk metals and alloys, (2) nanoparticles, (3) metal-nonmetal compounds and (4) carbon-based materials. In this Review, several theoretical investigations which have been realized in practice are highlighted in particular, such as metal-hydride nanocluster and carbon nitride supported copper complex for high-efficiency CO 2 reduction. It shows that the theoretical study can not only explain experimental phenomena but also predict improved candidates.