Urine deposition by grazing animals, with urea as the major constituent, is an important nitrogen (N) source in grassland ecosystems and may affect the production of three potent greenhouse gases (GHGs): nitrous oxide (N2O), methane (CH4) and carbon dioxide (CO2). However, owing to the greater chemical complexity of urine, the effect of urine on GHG emissions may differ from that of urea, and this difference has not been well reported. A short-term laboratory study was conducted to examine the influence of stored yak urine and urea at addition rates of 350 and 700 mg N m(-1) soil on GHG emissions at 20 and 60% water holding capacity (WHC) in an alpine steppe soil from the Qinghai-Tibetan Plateau. Compared with no N addition, urea led to greater increases in soil ammonium N (NH4+-N) concentrations, while yak urine resulted in a greater increase in soil dissolved organic carbon concentrations. Greater (P < 0.05) cumulative N2O and CH4 emissions occurred under 60% than 20% WHC, but soil moisture content did not affect CO2 emissions (P > 0.05). There were no statistically significant (P > 0.05) differences in N2O or CH4 emissions between the two N-addition rates, regardless of N type. The increases in N2O and CH4 emissions after N addition were greater in the urine than in the urea treatments, particularly under 60% WHC. The average CO2 emissions for the two soil moisture contents in the urine treatments increased with increasing N-addition rate, but the two urea treatments showed similar values to the control. The greater GHG emissions in the stored yak urine treatments were probably due to the existence of intrinsic organic carbon. Our results suggest that fresh yak urine may also possibly cause more GHG losses than urea under equivalent N application rates in alpine steppe soil on the Qinghai-Tibetan Plateau.