Sand-gravel mulch can effectively inhibit soil water evaporation, and hence promote crop growth. So far fewer researches have been reported on sand-gravel mulching than on plastic film or straw mulching, and little information is available about quantitative relationship between moisture and crop under sand-gravel mulch. To explore effects of sand-gravel mulching on soil water conservation and quantitative relationships between soil water evaporation and sand-gravel mulch, a soil water evaporation experiment, designed to have 5 mulching levels (0%, 25%, 50%, 75% and 100%), two irrigation levels (25 mm and 35 mm) and two soils (sand and lou soil), was carried out. Results show that (1) with the sand-gravel mulching increasing in coverage from 0% to 100%, the daily evaporation—time curve declined; daily evaporation was lower in plots irrigated with 25 mm of water than in the plots irrigated with 35 mm of water; and daily evaporation during the first 6 days decreased gradually with increasing sand-gravel mulching coverage, and after 7 days of evaporation, the decrease in daily evaporation tended to be lower, gradually; (2) the initial evaporation rate ( e_i ) ranged from 1.38 to 3.97 mm d^-1; it stepped down with increasing mulching coverage; e_w (steady evaporation rate) was significantly lower than e_i , varying in the range of 0.57 ~ 1.11 mm d^-1 as affected by sand-gravel mulching coverage; (3) with increasing sand-gravel mulching coverage, temporal variation curve of cumulated evaporation ( E_c ) declined and so did total evaporation ( E_t ); in the plots of bare soil, E_t accounted for 50% or so of the irrigation water (E), while E_t in the plots under 100% sand-gravel mulching did for about 35% of I only; and with increasing sand-gravel mulching coverage, evaporation to irrigation ratio decreased and evaporation difference (△E) increased; (4) The relationship between cumulative evaporation ( E_c ) and time (_t) fitted the theoretical relationship of Gardner, of which the parameter was highly related to mulching degree; thus a two parameter function was obtained for describing the relationships of cumulative evaporation and evaporation rate with mulching degree and time; and (5) at the end of the evaporation experiment, moisture contents in the soil profiles were much lower than the initial moisture content, regardless of mulching degree; but the moisture contents in the plots under mulch were generally high. To sum up, the effect of mulching on evaporation is apparent. The findings in the experiment may provide some important parameters in simulating soil water evaporation under gravel-sand mulch on a field scale.