Modelling soot formation and oxidation in diesel engines has been a long-standing challenge. In this study, soot particle characteristics in terms of particle dynamics, particle size and number density were modelled by integrating a multi-step phenomenological soot model into the KIVA-CHEMKIN CFD code for compression ignition engine combustion simulations. This semi-detailed soot model is dedicated to solving rate equations using sub-models to account for precursor formation, soot particle inception and coagulation as well as soot surface growth and oxidation. Soot growth and oxidation is inherently derived from the concentration of certain species found in the chemical reaction mechanism used, namely H, O-2, C2H2 and the nucleating polycyclic aromatic hydrocarbon (PAH). Acetylene is taken as the core precursor species in nucleating PAH, soot inception as well as acetylene-assisted soot surface growth. The integrated multi-step phenomenological soot model has been validated under gasoline and diesel dual-fuel engine conditions. The predicted histograms of soot particle number along with size distribution contribute towards the understanding of dominant factors that affect soot formation. The factors affecting soot particle under varying engine loads and fuel conditions were extensively investigated. Generally, the predicted soot particle size was larger for heavy-sooting conditions as compared to low-sooting conditions. For port-injected gasoline-dominated combustion, there is less soot being discharged. This might be attributed to two reasons. First, a more homogenous mixture is realized with less diesel fuel, thus effectively reducing the formation of soot precursors. The other reason is the intensive heat release which enhances the depletion of soot precursors, thereby impeding the growth of soot particles in terms of size and mass. (C) 2016 Elsevier Ltd. All rights reserved.