a b s t r a c tAlfalfa (Medicago sativa L.) is traditionally sown in field without ridge and furrow, in rain-fed semi-arid areas of Northwest China, and forage yields are normally low because of drought, water loss and soil erosion. A field study was conducted to determine (1) Monthly runoff efficiency and threshold rainfall for different ridge widths (30, 45 and 60 cm), and mulched with different materials (common plastic film, bio-degradable mulching film and manually compacted soil); and (2) the effects of different ridge-furrow ratios (30:60, 45:60 and 60:60) and ridge-mulching materials on soil water storage, topsoil temperature, forage yields, and water use efficiency (WUE) of Alfalfa in ridge-furrow rainwater harvesting (RFRWH) system, during 2 consecutive years: 2012 and 2013. Average monthly runoff efficiency over these 2 years, was 16%, 18%, 19%, 65%, 71%, 77%, 70%, 72% and 79% for SR 30 , SR 45 , SR 60 , BMR 30 , BMR 45 , BMR 60 , CMR 30 , CMR 45 and CMR 60 (SR, BMR and CMR were ridges with manually compacted soil, mulched with biodegradable mulch film and common plastic film, respectively, and subscripts 30, 45 and 60 refer to ridge widths (cm) all with 60 cm furrow width), respectively. For the same treatments, threshold rainfall was 4.3, 4.3, 4.2, 3.0, 2.9, 2.8, 2.9, 2.9 and 2.8 mm. The field experiment showed that ridge mulching materials and ridge widths had distinct effects on topsoil temperature at ridge tops, but not at furrow bottoms. Soil water storage in 140 cm soil depth at furrow bottoms increased with increasing ridge width and in the order of CMR ā BMR > SR > FP. Average annual total forage yields for FP, SR, BMR and CMR were 6276, 5642, 7563 and 7817 kg ha ā1 , respectively, over 2 years. The total precipitation in both years was higher than average of rainfall, leading to significant decreases of forage yields for SR, and significant increases of forage yields for BMR and CMR. Compared with FP, average forage yields decreased by 3%, 10% and 17% for SR 30 , SR 45 and SR 60 , and increased by 27%, 21%, 13%, 32%, 26% and 16% for BMR 30 , BMR 45 , BMR 60 , CMR 30 , CMR 45 and CMR 60 , respectively, over 2 years. WUE for SR, BMR and CMR was 1.69, 1.74 and 1.74 times greater than that for FP over 2 years, and increased with decreasing ridge width. Optimum furrow width was 35-36 cm for CMR and BMR in the RFRWH system with 60 cm width furrows for Alfalfa production in this region. Future study is needed to investigate optimum ridge-furrow ratio and suitable ridgemulching material for achieving the best environmental and economic benefit under different climatic conditions, soil types and plant species, and using bio-degradable mulching materials.