Complex networks play an important role in modern societies. Their failures, such as power grid blackouts, would lead to significant disruption of people's lives, industry and commercial activities, and result in massive economic losses. Reliable operation of these complex networks is an extremely challenging task. None of the complex network operations are fully automated; human-in-the-loop operation is critical. Given the complexity involved, there may be thousands of possible topological configurations at any given time. During an emergency, it is not uncommon for human operators to consider thousands of possible configurations in near real-time to choose the best option and operate the network effectively. In today's practice, network operation is largely based on experience with very limited real-time decision support, resulting in inadequate management of complex predictions and the inability to anticipate, recognize, and respond to situations caused by human errors, natural disasters, or cyber attacks. A systematic approach is needed to manage the complex operational paradigms and choose the best option in a near-real-time manner. This paper applies predictive analytics techniques to establish a decision support system for complex network operation management and help operators predict potential network failures and adapt the network in response to adverse situations. The resultant decision support system enables continuous monitoring of network performance and turns large amounts of data into actionable information. This paper presents examples with actual power grid data to demonstrate the capability of a proposed decision support system.