Fog computing has been proposed lately to offer services for the Internet of Things (IoT) thus adding many benefits for IoT applications. These benefits include support for low latency needs, mobility, location awareness, scalability, and efficient integration with other systems such as cloud computing. This paper investigates employing fog computing to enable multi-robot system (MRS) applications. It also discusses the potential services fog computing can offer MRS applications. Furthermore, the paper categorizes MRS applications highlighting the different fog support needed for each category. The paper classifies MRSs into five different categories based on the type of network used. This classification simplifies the process of understanding the issues involved in utilizing fog computing for different types of MRS applications. This classification also helps identify the most suitable fog computing architecture for each category. Then the paper proposes a service- oriented platform to support fog-enabled MRS applications and discusses its prototype implementation.