Modern CISs are becoming increasingly topologically interconnected and functionally interdependent. As a result, failure in one system may cause dependent components in other systems to fail, triggering cascading failures in networked CISs. Different CISs are heterogeneous in a variety of aspects, such as their topological characteristics and disaster resistance capacities. For instance, power grids are more susceptible than water supply systems to overload failure due to flow redistribution under disaster. Such systemic heterogeneity may significantly influence the failure propagation process across different CISs. However, despite the increasing volume of literature that examines failure propagation risks in networked CISs, few studies have accounted for systemic heterogeneity and its potential effects on cascading failures. The aim of this study is to assess the significance of such effects using one typical heterogeneity factor between the power and water supply systems. Firstly, a representative modeling approach of failure propagation, namely artificial flow based (AFB) approach, is selected through a thorough literature review. Secondly, two different artificial flow models (AFM) are developed using the AFB approach. Both models represent two interdependent, district-scale power and water systems, and they are distinguished by whether the systemic heterogeneity in susceptibility to overload damage is modeled by proper parameter settings. Lastly, both models are subjected to a simulated earthquake scenario, and three metrics are proposed to assess the overall responses of the two systems. The results from the two models are compared, which reveals that the magnitude of disaster impact of CISs would be notably overestimated when the systemic heterogeneity is not taken into consideration. The practical implications of the results are also discussed in the paper.