We survey some of the recent theoretical works about autonomous mobile robot systems and then discuss the possibility of extending robot models to analyze molecular computing systems. There are two types of robot systems appearing in the Distributed Computing literature. One is a self-reconfigurable system, which consists of a number of identical robotic modules that can connect to, disconnect from, and relocate relatively to adjacent modules. A behaviour of a self-reconfigurable system looks like the life game and we are interested in desigining an algorithm (local map) that makes the whole system behave in a consistent way. The other model is an autonomous mobile robot system, which consists of mobile robots with eye sensors as communication devices. The formation problem of a given geometrical pattern has been discussed extensively. Although those two systems have some common features and common goals, they were proposed independently and have been investigated separetely. After surveying some of works on those robot systems, we introduce some works in molecular computing. We explain some of the examples in which those robot systems appear naturally. We then argue some problems that arise from the molecular computing applications. A key coincidence between the robot models and the molecular computing is that they are systems composed of anonymous modules.