Soot modeling in diesel engines has evolved over four decades from simple empirical relations to detailed kinetics involving soot precursors and complex particle dynamics. Although numerical models are well established for the prediction of diesel soot emissions, most of the models are validated only for the mass of soot emissions. Validating particle size and number density with engine data has rarely been attempted due to the lack of experimental data and because particle number density was not included in emission norms until recently for EURO 6. The ability to predict particle dynamics for parametric variations provides a knowledge of evolution of particle size distribution inside an engine and helps to design better combustion systems. The current work aims at modeling the soot particle size distribution inside a heavy duty and a light duty diesel engine and validating the results for a variation of injection pressure. A sectional soot model has been applied for closed cycle combustion simulations in CONVERGE, a 3D CFD code. Transport equations for soot in each section are solved along with source terms for nucleation, surface reactions and coagulation coupled with gas phase kinetics representing fuel oxidation and soot precursor chemistry. The effect of varying critical parameters involved in the soot model has been discussed in detail. Based on the aerosol regimes in diesel engines, the collision frequency estimation based on transition regime assumption has been found to be inadequate and an enhancement by a factor of hundred has been found to provide a good agreement between numerical predictions and experimental data for soot particle size distribution (PSD) at exhaust. Evolution of in-cylinder soot PSD has been explained based on soot particle dynamics. Numerical predictions show encouraging qualitative trends for soot PSD at the exhaust condition. This improved method of prediction of soot mass and PSD helps to design engines to meet EURO 6 emission norms.