This study aims to determine whether Hass avocado irrigation can be triggered based on the surface soil water content (SSWC). To address this question, the soil water dynamics from three Hass avocado orchard plots located in Valle del Cauca (Colombia) was simulated using a model provided by Hydrus-1D software, which was calibrated through the genetic algorithm NSGA-II and validated using the soil matric potential measured at several depths at nine monitoring stations installed in the three plots. The influence of each superficial (0–5, 5–10, and 10–15 cm) and deeper (15–30 and 30–60 cm) available water (AW) computed from the simulated moisture on the SWB at 0–60 cm was estimated from the artificial neural network (ANN) trained weights. The most influential depth range was used to predict the soil water balance at 0–60 cm using ANN. For validation, the RMSE slightly increased regarding calibration, varying from 1.73 to 8.20, while the R2 value varied from 0.61 to 0.89 (P < 0.001 for all cases). The AW at 5–10 cm depth had a significant influence on SWB with an average relevance index of 2.87 (Wilcoxon signed-rank test P ≤ 0.05) for Laurentina farm. The AW at 0–5 cm depth had not significant influence on SWB with an average relevance index of 1.34 (independent group) and 0.97 (P < 0.05) for Laurentina and Poncena, respectively. The ANN model predicted the SWB with a RMSE no bigger than 13.76 mm. In conclusion, the SSWC at a depth of 5–10 cm can be used as an indicator for scheduling Hass avocado irrigation.