In recent years, Distributed Autonomous Robotic System (DARS) becomes a major concern in developing robots of many projects. The use of DARS to accomplish tasks within a complex environment with limited resources requires strategies that can leverage the inherent redundancies within these systems. In fact, different DARS strategies produce differents DARS architectures, and the practical level of DARS design is reflected by reliability of DARS architectures. However, there are still some challenges, which contain communication, control, perception and energy management, in reliability evaluation of DARS architectures design. Therefore, considering these challenges, reliability evaluation model study of DARS architectures is needed in engineering. In order to evaluate reliability of DARS architectures systematically, and guide DARS architectures design effectively, this paper provides a reliability evaluation model of DARS architectures. There are five factors in this model, such as communication ability, adaptation ability, information integration, task allocation, and destination achievement. This model will evaluate these factors separately and provide a combined evaluation result together. Additionally, this paper explains a case to demonstrate how this model could be used in projects. The case indicates that the model could take advantage of reliability factors together, and gives a reasonable reliability evaluation in engineering. In addition, this paper makes a discussion in model improvement for different architectures reliability evaluation. Finally, the advantages of this model and the future work will be summarized.