We propose an adaptive cyber security monitoring system that integrates a number of component techniques to collect time-series situation information, perform intrusion detection, keep track of event evolution, and characterize and identify security events so corresponding defense actions can be taken in a timely and effective manner. Particularly, we employ a decision fusion algorithm with analytically proven performance guarantee for intrusion detection based on local votes from distributed sensors. Different from the traditional rule-based pattern matching technique, security events in the proposed system are represented in a graphical form of correlation networks using random matrix theory and identified through the computation of network similarity measurement. Extensive simulation results on event identification illustrate the efficacy of the proposed system. Index Terms-Cyber security, decision fusion, event correlation, random matrix theory.