Changes in precipitation are affecting the dynamics of C, N, and P in ecosystems, which are coupled within plants, microbes, and soils. Therefore, studying the responses of C:N:P stoichiometry in plants, microbes, and soils to changing precipitation levels are significant for elucidating how biogeochemical cycling responds to global climate change. Based on a two-year simulated experiment of precipitation change in a desert steppe in the Ningxia region, which was conducted from May 2014 to August 2015 and included five treatments (50% reduction in precipitation, 30% reduction in precipitation, natural precipitation, 30% increase in precipitation, and 50% increase in precipitation), we measured the levels of C, N, and P in plants, microbes, and soils, and explored the C:N:P stoichiometry in these three components of plant-soil systems. Meanwhile, plant communities were investigated and soil water contents, pH levels, and temperatures were monitored. Through the correlation analyses between soil indices and plant and microbe indices, the indications of soil C:N:P stoichiometry for the growth and nutrient conservation of plants (uptake and resorption) and the growth of microbes were estimated, respectively. The results showed that 1) two years of precipitation treatments altered the C:N:P stoichiometry in green leaves of plants and the influences were different among species. Generally, a 50% reduction in precipitation increased the N and P uptakes in green leaves of Lespedeza potanimill and P uptake in green leaves of Artemisia scoparia, whereas both 30% and 50% increases in precipitation decreased the N uptake in green leaves of A. scoparia. Both 30% and 50% increases in precipitation increased the C:N in green leaves of A. scoparia, whereas only a 30% increase in precipitation increased the C:N in green leaves of Sophora alopecuroides. Both 30% and 50% increases in precipitation reduced the N:P in green leaves of A. scoparia, whereas only a 30% increase of precipitation decreased the N:P in green leaves of Pennisetum centrasiaticum. In contrast, two years of precipitation treatments had lesser effects on the C:N:P stoichiometry in senescing leaves of the studied species. 2) Reduced precipitation decreased microbial biomass C, N, and C:N, whereas increased precipitation stimulated an increased microbial biomass accumulation. However, a 50% increase in precipitation inhibited the C accumulation of microbes and thus reduced the C:N. 3) Two years of precipitation treatments had little effect on the soil C:N:P stoichiometry (except soil organic C and C:N). Increased precipitation improved the soil water availability, and thus stimulated the growth of plants and microbes. 4) During 2014-2015, a relatively stable soil C:N:P stoichiometry could not indicate the nutrient limitations for plant growth and microbe reproduction. The self-adjusted nutrient conservation strategy, specifically in increased leaf nutrient uptake when supplied with low precipitation, and enhanced leaf nutrient resorption when supplied with high precipitation, probably explains the high adaptation ability of L. potaninii in response to precipitation change.