The effect of different forest conservation policies on water provision has been poorly investigated due to a lack of an integrative methodological framework that enables its quantification. We developed a method for assessing the effects of forest conservation policies on water provision for rural inhabitants, based on a land-use model coupled with an eco-hydrological model. We used as a case study the Lumaco catchment, Chile, a territory dominated by native forests (NF) and non-native tree farms, with an extended dry period where nearly 12,600 people of rural communities get drinking water through water trucks. We analyzed three land-use policy scenarios: i) a baseline scenario based on historical land-cover maps; ii) a NF Recovery and Protection (NFRP) scenario, based on an earlier implementation of the first NF Recovery and Forestry Development bill; and iii) a Pristine (PR) scenario, based on potential vegetation belts; the latter two based on Dyna CLUE, and simulated between 1990 and 2015. Impacts on water provision from each scenario were computed with SWAT. The NFRP scenario resulted in an increase of 6974 ha of NF regarding the baseline situation, and the PR scenario showed an increase of 26,939 ha of NF. Despite large differences in NF areas, slight increases in inflows (Q) were found between the NFRP and the PR scenarios, with relative differences with respect to the baseline of 0.3% and 2.5% for NFRP and PR, respectively. Notwithstanding, these small differences in the NFRP scenario, they become larger if we analyze the cumulative values during the dry season only (December, January, and February), where they reach 1.1% in a normal year and 3.1% in a dry year. Flows increases were transformed into water truck costs resulting in up to 441,876 USD (monthly) of fiscal spending that could be avoided during a dry period.