Bio-derived fuels are drawing more and more attention in the internal combustion engine (ICE) research field in recent years. Those interests in use of renewable biofuels in ICE applications derive from energy security issues and, more importantly, from environment pollutant emissions concerns. High fidelity numerical study of engine combustion requires advanced computational fluid dynamics (CFD) to be coupled with detailed chemical kinetic models. This task becomes extremely challenging if real fuels are taken into account, as they include a mixture of hundreds of different hydrocarbons, which prohibitively increases computational cost. Therefore, along with employing surrogate fuel models, reduction of detailed kinetic models for multidimensional engine applications is preferred. In the present work, a reduced mechanism was developed for primary reference fuel (PRF) using the directed relation graph (DRG) approach. The mechanism was generated from an existing detailed mechanism. The adjustment of reaction rate constants of selected reactions was performed and the present reduced mechanism was validated against experiments in terms of ignition delay times, flame speed and HCCI combustion. Employing similar procedures, reduced reaction mechanisms for ethanol and butanol were generated and incorporated into the PRF mechanism to be able to model multi-component gasoline-biofuel combustion in SI engines. The results show that the present reduced mechanism demonstrates reliable performance in combustion predictions, as well as significant improvement of computational efficiency in multi-dimensional CFD simulations.