Water system innovations such as rainwater harvesting involve abstraction of water in the upper catchments. Increasing adoption of rainwater harvesting in the riparian catchments could have hydrological impacts on downstream flows in the river basin, but it is assumed to have overall gains and synergies when efficient use of rainwater is optimized at farm-level. This paper examines the main determinants of adoption of water system innovations with specific emphasis on the intensity of adoption and adoption lag, using a cross-sectional sample of 234 farmers in the Makanya watershed. Censored Tobit models were used to estimate the coefficients of intensity of adoption and adoption lag of water system innovations. Group networking, years spent in formal education, age of respondent, location and agricultural information pathways were found to be major determinants of intensity of adoption at farm-level. It was also found that intensity of adoption and frequency of attendance to collective action are strong determinants of adoption lag of water system innovation in Makanya watershed. Empirical knowledge of the determinants of adoption of water system innovations is critical for an effective scaling out of best practices of water harvesting in the Basin.