Water conservation technology is widely used in developing countries to avoid the adverse effects of water scarcity. Data obtained from a survey of 385 Nepalese households were analyzed by using parametric, semiparametric and nonparametric models. We used a Poisson quasi-likelihood method to estimate a parametric model, and local-linear least squares cross-validation bandwidth selection method for a semiparametric model. A test developed by Hsiao et al. (2007) was employed to compare parametric model to semiparametric model, and we found the semiparametric model more appropriate than a parametric model in the analysis of the intensity of water conservation technology. Semiparametric model allows the entry of variables parametrically and nonparametrically to describe linear and nonlinear effects. Results show that the farmers having agriculture as the main source of income and sufficient food availability in the previous year are positively and significantly associated with the intensity of technology adoption. These results signify the needs for appropriate policies and programs to enhance farm production and income that would encourage farmers to adopt water conservation technologies for sustainable agriculture.