In this work we developed a RAFT polymerization model taking into account the main reactions of the experimental RAFT process and implemented that model in dissipative particle dynamics (DPD). With the help of a kinetic model based on the same reaction routine, we investigated the question of how to simulate realistic reactions using such models. We showed that a simultaneous M-fold increase of the initiation probability p_i and an M-fold decrease of the termination probability p_t does not result in significant changes in the chain length distribution. If the RAFT/initiator ratio is large, a simplified model with no termination and immediate radical formation can be used with good enough accuracy. After that we directly compared the reaction behavior within the kinetic model and DPD. We showed that steric restrictions, which were not present in the kinetic model, can introduce noticeable changes in the system behavior. Finally, we studied the influence of the incompatibility on the RAFT polymerization process on an example of classical implementation of polymerization-induced self-assembly (PISA). We showed that in systems with incompatible species the number of activation–deactivation cycles does not always reflect the dispersity of the resulting chain ensemble. Moreover, we demonstrated that specifically the incompatibility between the RAFT end group and other species can have a large effect on the polymerization results.