<p>Accurate estimation of evapotranspiration (ET) and its components is challenging in semi-arid rain-fed farmlands because of the complicated interactions among vegetation, soil, and water conditions. In this study, three stomatal conductance models were used to refine the canopy stomatal conductance in the Shuttleworth-Wallace (SW) model. The model was calibrated and validated using eddy covariance measurements and micro-lysimeter data. The results demonstrated that coupling the Ball-Berry (BB) stomatal conductance model with the SW model yielded the highest simulation accuracy, rendering it suitable for estimating ET and its components in semi-arid rain-fed maize fields. The model incorporating soil water content exhibited notable sensitivity to variations in soil water content, highlighting its significant impact on semi-arid rain-fed maize ecosystems. Throughout the 2015–2018 growing seasons, daily ET, E, and T varied in the range of 0.055–6.5&#xa0;mm/d, 0.036–1.96&#xa0;mm/d and 0.004–5.64&#xa0;mm/d, respectively. E/ET varied between 0.29 and 0.32, with an average of 0.31. Stomatal conductance (g_s) had a significant effect on E/ET on a daily timescale, and E/ET decreased as g_s increased. The monthly E/ET was primarily influenced by air temperature (<i>Ta</i>) and leaf area index (<i>LAI</i>), with <i>Ta</i> indirectly affecting E/ET through its effect on <i>LAI</i>.</p> Graphical abstract <p></p>