Aims Soil respiration of the lands covered by biocrusts is an important component in the carbon cycle of arid, semi-arid and dry-subhumid ecosystems (drylands hereafter), and one of the key processes in the carbon cycle of drylands. However, the responses of the rate of soil respiration with biocrusts to water and temperature are uncertain in the investigations of the effects of experimental warming and precipitation patterns on CO_2 fluxes in biocrust dominated ecosystems. The objectives of this study were to investigate the relationships of carbon release from the biocrust-soil systems with water and temperature in drylands. Methods Intact soil columns with two types of biocrusts, including moss and algae-lichen crusts, were collected in a natural vegetation area in the southeastern fringe of the Tengger Desert. Open top chambers were used to simulate climate warming, and the soil respiration rate was measured under warming and non-warming treatments using an automated soil respiration system (LI-8150). Important findings Over the whole observational period (from April 2016 to July 2016), soil respiration rates varied from -0.16 to 4.69 mumol·m~(-2)·s~(-1) for the moss crust-covered soils and from -0.21 to 5.72 mumol·m~(-2)·s~(-1) for the algae-lichen crust-covered soils, respectively, under different rainfall events (the precipitations between 0.3-30.0 mm). The mean soil respiration rate of the moss crust-covered soils is 1.09 mumol·m~(-2)·s~(-1), which is higher than that of the algae-lichen crust-covered soils of 0.94 mumol·m~(-2)·s~(-1). The soil respiration rate of the two types of biocrust-covered soils showed different dynamics and spatial heterogeneities with rainfall events, and were positively correlated with precipitation. The mean soil respiration rate of the biocrust-covered soils without warming was 1.24 mumol·m~(-2)·s~(-1), significantly higher than that with warming treatments of 0.79 mumol·m~(-2)·s~(-1) (p < 0.05). By increasing the evaporation of soil moisture, the simulated warming impeded soil respiration. In most cases, soil temperature and soil respiration rate displayed a similar single-peak curve during the diel cycle. Our results show an approximately two hours lag between soil temperature at 5 cm depth and the soil respiration rate of the biocrust-covered soils during the diel cycle.