The high pressure injectors used in direct injection Diesel engines introduce major perturbations in the air flow field inside the combustion chamber leading to strongly strained and turbulent flow. This fuel/air mixing process plays a critical role in enhancing self-ignition. However, in most Diesel combustion models, the interaction between turbulent mixing and self-ignition is not directly taken into account. Typically, the calculated average self-ignition combustion rates are pseudo laminar reaction rates based on simplified kinetic mechanisms. The mean values of the reaction rate are determined as a function of the mean values of the reactant concentrations and temperature. But due to the high non linearity of the reaction rate during self-ignition, this assumption is not valid. A turbulent self-ignition model developed from direct numerical simulations is presented. It uses a presumed probability density function approach that takes into account the history of the turbulent mixing. The self-ignition model is coupled to a flamelet approach for the later stages of combustion at high temperature. The model is applied to the calculation of a constant volume combustion chamber configuration filled with air in Diesel thermodynamic conditions where a gaseous fuel is injected at different velocities. To validate the approach, experiments are made on self-ignition and combustion of high pressure methane jets injected inside a high pressure combustion cell. The first results comparing calculated non reactive jet profiles and experimental visualizations have shown good agreement. Comparisons of self-ignition delays and combustion visualizations have also been conducted showing that the turbulent self-ignition and combustion model reproduces well the first ignition locations and flame propagation. The trends in the self-ignition delays when the flow conditions have changed have also been reproduced by the calculations.