Water distribution systems, electrical power transmission systems, and other civil infrastructure systems are essential to the smooth and stable operation of regional economies. Since the functions of such infrastructure systems often are inter-dependent, the systems sometimes suffer unforeseen functional disruptions. For example, the widespread power outage due to the malfunction of an electric power substation, which occurred in the northeastern United States and parts of Canada in August 2003, interrupted the supply of water to several communities, leading to inconvenience and economic losses. The sequence of such failures leading to widespread outages is referred to as a cascading failure. Assessing the vulnerability of communities to natural and man-made hazards should take the possibility of such failures into account. In seismic risk assessment, the risk to a facility or a building is generally specified by one of two basic approaches: through a probabilistic seismic hazard analysis (PSHA) and a stipulated scenario earthquake (SE). A PSHA has been widely accepted as a basis for design and evaluation of individual buildings, bridges and other facilities. However, the vulnerability assessment of distributed infrastructure facilities requires a model of spatial intensity of earthquake ground motion. Since the ground motions from a PSHA represent an aggregation of earthquakes, they cannot model the spatial variation in intensity. On the other hand, when a SE-based analysis is used, the spatial correlation of seismic intensities must be properly evaluated. This study presents a new methodology for evaluating the functionality of an infrastructure system situated in a region of moderate seismicity considering functional interactions among the systems in the network, cascading failure, and spatial correlation of ground motion. The functional interactions among facilities in the systems are modeled by fault trees, and the impact of cascading failures on serviceability of a networked system is computed by a procedure from the field of operations research known as a shortest path algorithm. The upper and lower bound solutions to spatial correlation of seismic intensities over a region are obtained.