Water treatment plants (WTPs) are characterised as complex configurations of repairable and deteriorating components. Previous studies have mainly focused on the average or steady-state availability of such systems while ignoring inherent characteristics like degradation. The current research proposes a two-level hierarchical model for long-term availability analysis of WTPs. To do so, at the component level, a condition-based technique (semi-Markov) or a failure-based technique (non-homogeneous Poisson process) is proposed based on the type and amount of available data while at the system level a reliability block diagram can be used to combine the component-level availabilities. The application of the methodology has been demonstrated on a real case study in the Netherlands.