There is a growing interest in the understanding of the dynamics of sap flow and the transpiration process in plants. In this study, Xylem sap flux was monitored in a 4-year-old cashew (Anacardium occidentale) orchard for eight months (2–4 weeks per month), covering both wet and dry seasons of 2002 at Ejura in Ghana. Sap flux (Fd) was related diurnally with vapor pressure deficit (De), incoming solar radiation (Rs), and estimated evaporative demand (Eo) on a monthly and a seasonal basis. Cross-correlation analysis was used to estimate time lag (τ) between diurnal pattern of Fd and Rs, De or Eo. The maximum correlation coefficient (rmax) from cross-correlation of Fd with Rs (range = 0.85–0.93) was often slightly higher than rmax of Fd and De (range = 0.83–0.92), indicating that the diurnal Fd was more dependent on Rs than De for most part of the year. Generally, τ of Fd with Rs ranged from 0–60 min, 0–30 min with Eo, and led De from 0–180 min during the year. Student t-test showed that average diurnal Fd was not statistically different, from month to month, all year round, suggesting that trees were able to source for required water even during February and March, which are the peak dry months in West Africa. Regressions between Fd and De or Rs showed that De exerted more control on tree water use during the wet (June, July, October) than dry (January, February, March) or transition (November, December) periods. The good relation of Fd with the climatic parameter was exploited to calibrate an approximation function to predict seasonal water use to a very high degree of accuracy. A simple mono-climatic model, combining the parabolic response of Fd to Rs and the linear response to De, was highly successful (F = 9118.5, P < 0.0001, r2 = 0.82, n = 3984) in predicting sap flux in the orchard. This result suggested that sap flow data might be used to estimate stand transpiration conveniently at about hourly time steps to avoid the usual time-lag problem associated with the use of sap flow data for transpiration estimates.