NASA is currently evaluating the benefits of transitioning to a highly reconfigurable network of arrayed dish antennas to support an increasing number of deep space missions. The next-generation Deep Space Network (NG-DSN), as currently conceptualized, would require extensive automation to reduce operations cost and handle the increased complexity associated with monitoring and controlling the larger number of antennas. This paper presents a prototype operations architecture for the proposed NG-DSN that is fundamentally based on three concepts: physical state variables of the system to be controlled, expressions of operational intent for those state variables (“goals”), and models describing the behavior of these state variables and their interactions. These concepts shape the software design of an automated control system, the model-based systems engineering analysis that feeds this design, and the human operator interface to the control system. This control system provides for automation of capabilities such as resource allocation and fault recovery (both localized and system-wide). This paper describes the development and demonstration of the control system on prototype antenna array hardware at the Jet Propulsion Laboratory.