Soil respiration rates under five land use patterns in the water-wind erosion crisscross region of the Loess Plateau were measured by the closed-chamber IRGA method during the growing season of 2007. Differences in soil respiration under the five land use patterns and the relationships between soil respiration and soil temperature and soil water contents were analyzed. The results showed that seasonal changes of soil respiration presented single peak curves, and that of air temperature followed a similar trend, which had the highest values in July and August. The order of average soil respiration rates under the land use types during the growing seasons was: bunge needlegrass land > alfalfa land > Korshinsk Peashrub land > cropland > sand willow land. The soil respiration rate of the grassland was significantly higher than that of the crop and shrub lands. Except for sand willow and alfalfa lands, soil respiration rates were better correlated with soil temperature at 10 cm depth than with other temperature indexes. According to Q10 values (temperature sensitive index), soil respiration under cropland was the most sensitive to temperature (Q10=2.20) and Q10 of the other land use patterns, except for sand willow land, were around 2.0, which is close to the global average Q10 value. Estimating the soil respiration flux using the Van′t Hoff model gave CO2 effluxes from alfalfa land, bunge needlegrass land, Korshinsk Peashrub land, crop land, and sand willow land, during the growth period, of 259 gC·m-2, 236 gC·m-2, 226 gC·m-2, 170 gC·m-2, and 94 gC·m-2, respectively. Soil moisture did not significantly affect soil respiration under crop and sand willow lands. A two-variable (soil temperature and soil moisture) soil respiration model best explained the variance of soil respiration under alfalfa land, bunge needlegrass land, and Korshinsk Peashrub land.