Understanding and modeling soot particle dynamics in combustion systems is a key issue in the development of low emission engines. In engines, soot particles are formed as a result of complex hydrocarbon chemistry and are subject to a turbulent flow field which controls ultimately the yield of soot particles. In this work, we will detail the strategies used to model the various chemical and physical processes encountered both in laminar and turbulent flames. More precisely, we will consider the impact of the chemistry on the inception of the first soot particles, the geometrical and statistical representation of fractal aggregates, the oxidation and fragmentation of particles under lean conditions, and finally the turbulent transport of soot in complex unsteady flows. For each of these cases, we will compare our results with experimental measurements and discuss the differences. Finally, we will show preliminary results for the first Large Eddy Simulation of a sooting turbulent jet diffusion flame with detailed chemical and soot models. This last simulation highlights the challenges in modeling soot evolution in turbulent flames due to the nonlinear interactions between the particles and the gas-phase turbulent combustion processes.