Over last two decades, micro-rainwater harvesting technology has been widely used in dryland agricultural practice, but the effects of the design of this technique and the variability of regional rainfall on the storage, migration and utilization of rainwater in soil is very rare. Field experiments were conducted using the spring oat (Avena sativa L.) variety, Dingyou 3 as test material, at Dingxi Agricultural Meterological Experimental Station of Gansu Province in 2009 and 2010. Despite the total rainfall amount for the growing period of oat being basically consistent in 2009 and 2010, there was a significant difference in the temporal distribution of rainfall in the two growing seasons. In 2009, the rainfall distribution was similar to the long-term trend in rainfall (consisting of 33.1% of total rainfall in the first half and 66.9% in the second half of whole growth period). Conversely, the year 2010 turned to be characterized by "more in the first half and less in the second half" (accounting for 46.4% and 53.6% of total rainfall of whole growth period, respectively). The coefficients of variation of rainfall distribution were 1.0 and 0.8 in 2009 to 2010, respectively. In this study, eight treatments of micro-rainwater harvesting were designed, including: (1) the flat bare plot without cultivation (CK group 1), (2) flat plot with cultivation (CK group 2), and (3) six treatments with ridge-furrow and plastic mulch (3 width ratios of ridge to furrow (2 ∶ 1, 1 ∶ 1 and 2 ∶ 3), and two mulching patterns (plastic mulch and no mulch). Water storage and consumption in the soil profile, water use efficiency and yield-related indicators were collected from all the treatment groups in both years. The data from the two years showed that the leaf area index, water use efficiency and yield were significantly higher in the three treatments with film mulching on the ridge than those of other treatments. The patterns of film mulching on the ridge were found to improve water conversion efficiency from rainwater into soil and crop, and induced a rapid recovery of the water stored in soil in the later growth period. With respect at the width ratios of ridge to furrow, the parameters mentioned above were the highest in the treatment groups with 1 ∶ 1 width ratios of ridge to furrow in all treatments. Furthermore, the stored water in the treatments of 1 ∶ 1 or 2 ∶ 1 width ratios from ridge to furrow was significantly greater than that of the 2 ∶ 3 ratio treatment groups. Importantly, analysis of the performance in the two years with different temporal variability in rainfall showed that water use efficiency, grain yield and leaf area index of oat were all significantly higher in 2010 than those in 2009 (P<0.05), and the total volume of water stored in 0-130 cm soil profile were also higher in 2010 than 2009. Our results suggest that the "forward transfer" of rainfall distribution was conducive to optimizing the soil ecological processes in the case of similar total rainfall amount. Our study provides a foundation for future understanding on the effects of micro-rainwater harvesting pattern and rainfall change on ecological processes of dryland agriculture, and also provides a new strategy for phenological management in rain-fed farming regions.