OBJECTIVE: Access to improved water sources is rapidly expanding in rural central Vietnam. We examined one NGO-led piped water supply programme to assess the drinking water quality and health impacts of piped water systems where access to 'improved' water sources is already good. METHODS: This longitudinal, prospective cohort study followed 300 households in seven project areas in Da Nang province, Vietnam: 224 households who paid for an on-plot piped water connection and 76 control households from the same areas relying primarily on 'improved' water sources outside the home. The 4-month study was intended to measure the impact of the NGO-led water programmes on households' drinking water quality and health and to evaluate system performance. RESULTS: We found that: (i) households connected to a piped water supply had consistently better drinking water quality than those relying on other sources, including 'improved' sources and (ii) connected households experienced less diarrhoea than households without a piped water connection (adjusted longitudinal prevalence ratio: 0.57 (95% CI 0.39-0.86, P = 0.006) and households using an 'improved' source not piped to the plot: (adjusted longitudinal prevalence ratio: 0.59 (95% CI 0.39-0.91, P = 0.018). CONCLUSIONS: Our results suggest that on-plot water service yields benefits over other sources that are considered 'improved' by the WHO/UNICEF Joint Monitoring Programme.