Wireless sensor networks (WSNs) consist of spatially distributed sensor nodes to monitor physical or environmental conditions. These sensor nodes are typically battery-powered sensor nodes (BPSNs) and can rarely meet design goals of long network lifetime and high reliability. Energy-harvesting sensor nodes (EHSNs) that convert different types of energy to electrical energy are an alternative type of sensor nodes with a long lifetime but a high cost. Combining BPSNs and EHSNs has potential to deal with the conflicting design goals of long lifetime and reasonably low cost. In this study, the authors make new contributions by modelling a heterogeneous WSN consisting of both BPSNs and EHSNs and proposing a comprehensive cost function-based routing approach that integrates end-to-end path reliability, cost and energy consumption for providing satisfactory quality of service to applications running on hybrid WSNs. Another contribution made in this work is the optimal deployment of EHSNs using a reliability-importance-analysis-based method to improve the end-to-end path reliability within hybrid WSNs.