Hydrological processes in arid/semi-arid agricultural watersheds are challenging to simulate accurately due to the strong interactions within the hydrological cycle and the influences of agricultural activities. This paper introduces a new mechanistic, distributed hydrological model (named SurfMOD), developed to effectively simulate the dynamics of water and salt in the vadose zone, groundwater, surface water, plant, and their interactions over time and space. The processes are simulated based on the numerical solution of partial differential equations and a practical spatiotemporal coupling method is proposed to ensure the computational efficiency. In addition, SurMOD provides functions to consider various agricultural practices (e.g., surface mulching, soil bunds, canal seepage, and multi-order drainage system). SurfMOD was tested using observed data from typical arid/semi-arid agricultural areas in the upper Yellow River basin: an experimental site (Yangchang canal command area, YCA) during 2012-2013; and a region-scale area (Jiyuan Irrigation System, Jiyuan) in 2021. The results showed excellent agreement between the simulations and observed data, with coefficients of determination (R2) exceeding 0.44, 0.39, 0.38, and 0.87 for soil water content and salinity concentration, groundwater level, ditch drainage rate, and leaf area index, respectively. The Nash-Sutcliffe Efficiency was satisfactory, and the root mean square error was acceptably low in both case studies. Furthermore, the irrigation scenario tests and comparisons to previous studies demonstrate that SurfMOD offers significant advantages in both simulation accuracy and computational efficiency. Overall, SurfMOD enhances the capability of hydrological models to simulate coupled processes of complete water cycling and plant growth in arid/semi-arid agricultural watersheds. It could become a valuable tool for the quantitative analysis of arid/semi-arid agroecosystems, scenario forecasting, and the efficient management of water use and salinity control on a regional scale.