Ensuring the reliability of water treatment systems is a critical engineering challenge for sustainable urban infrastructure in West Africa. Existing assessments often lack longitudinal rigour and robust statistical frameworks, hindering effective maintenance planning and investment. This paper aims to methodologically evaluate the operational reliability of major water treatment facilities and to develop a panel-data econometric model for estimating and predicting system failure rates. A balanced panel dataset was constructed from technical performance records of 24 urban treatment plants. Reliability was quantified via a failure rate index. The core estimation employs a two-way fixed effects model: $FailureRate_{it} = \alpha + \beta_1 Age_{it} + \beta_2 Maintenance_{it} + \mu_i + \lambda_t + \epsilon_{it}$, with inference based on cluster-robust standard errors. Plant age and preventative maintenance expenditure were statistically significant predictors of reliability. A 10% increase in scheduled maintenance spend was associated with a 3.2 percentage point reduction in the annual failure rate (95% CI: 1.8 to 4.6). The model forecasts a gradual decline in aggregate system reliability without intervention. The panel-data approach provides a robust methodological framework for quantifying water treatment system reliability, revealing significant, actionable drivers of performance. Infrastructure asset management should integrate panel-data modelling for predictive maintenance scheduling. Policy must prioritise ring-fenced funding for preventative maintenance to counteract ageing infrastructure effects. Infrastructure reliability, panel data, fixed effects model, water treatment, asset management, predictive maintenance This paper provides a novel application of panel-data econometrics to the longitudinal analysis of water treatment system performance, generating a predictive tool for infrastructure management.