In this work we propose a method of generating alternatives for the evolution of processes and systems using a simulation modeling system in accordance with the agent-based approach. Use of this method allows to automate the process of analyzing real world processes and complex systems for their further management. The paper presents a description of proposed method design that includes the modeling notation and the algorithm for generating evolution alternatives. A graph model is used to describe the agent’s behavior, and the technique of grouping its possible states into classes is used to improve the efficiency of modeling and to make further analysis of the results more convenient. The algorithm for generating evolution alternatives supports various strategies for selecting possible alternatives, including the probabilistic model. The proposed method and a set of related components are implemented at the software level in the form of a simulation modeling system. With the help of this system, the study of the creation of alternatives for the evolution of sample processes with high risks in case of making an incorrect decision has been carried out. The use of the proposed method reduces the time required to create all possible alternatives of sample processes evolution compared to manual enumeration by 400–7044 times (depending on the corresponding model size and complexity).