To determine the effects of soil moisture dynamics on hydrological and ecological processes in a ridge-furrow rainwater harvesting system in a semiarid area, a field experiment with a randomized complete block design was conducted, which evaluated the effects of different mulching materials (common plastic film, biodegradable mulch film and manually compacted soil) and different ridge-furrow ratios (30:60, 45:60 and 60:60 (cm:cm)) on soil moisture dynamics in the root zone of oats through model simulation and using the continuously monitored data of soil moisture during two consecutive oats growing years, 2012 and 2013, and daily precipitation from 2000 to 2015 at the Dingxi Arid Meteorology and Ecological Environment Experimental Station. The simulation was carried out using a stochastic model (Laio model) for soil moisture dynamics. The probabilistic density function (PDF) of soil moisture at a point scale was estimated using the Laio model, and the parameter sensitivity of the model was analyzed in this study. The results indicated that the precipitation distribution was irregular, and the precipitation events happening between May and October accounted for 66.6% of total annual precipitation events. Approximately 85.32% of annual rainfall received from rainfall higher than 10 mm. Annual precipitations had an increase trend in the past 16 years in this region. The threshold rainfall to produce runoff was 1.35, 0.95, and 5.31 mm for BMR (ridges mulched with biodegradable mulch film), CMR (ridges mulched with common plastic film), and SR (ridges with manually compacted soil), respectively, whereas the average runoff efficiency for the same treatments was 87.892%, 94.203%, and 27.488%. Soil water content for BMR and CMR was significantly higher than that for SR, which was significantly higher than that for traditional planting without ridges and furrows. The soil water content for all treatments followed a normal distribution. The curve characteristics (the curve peak value, position of the peak, 90% confidence interval) and the digital features (the expectations and variances) of the soil moisture PDF simulated by the model have a good consistence with observed values, and the consistency measure (CM) was higher than 0.5. The rainwater coupled with runoff in furrows could be considered a random event. The Laio model could be used for estimating the soil moisture PDF in ridge-furrow rainwater harvesting systems, and this could provide an insight for soil moisture management with a high use efficiency in semiarid regions.