The Qinghai-Tibetan Plateau (QTP) is sensitive to global climate change. This is because it is characterized by irregular rainfall and freeze–thaw cycles resulting from its high elevation and low temperature. Greenhouse gases (GHGs) mainly contribute to the warming of the QTP, but few studies have investigated and compared the effects of irregular rainfall and freeze–thaw cycles on GHGs. In this study, we conducted a laboratory experiment under four types of freeze–thaw treatments with three soil water content levels to simulate the irregular freeze–thaw and rainfall conditions. The results showed that both the soil water content and freeze–thaw treatment influenced the soil properties, soil enzyme activities, and the microbial biomass; however, the freeze–thaw treatment had significantly higher influences on GHG fluxes than soil water content. In order to explore other biotic and abiotic factors in an attempt to establish the main factor in determining GHG fluxes, a variation partition analysis was conducted. The results revealed that freeze–thaw treatments were the strongest individual factors in predicting the variance in N2O and CO2 fluxes, and the pH, which was only significantly affected by freeze–thaw treatment, was the strongest individual factor in predicting CH4 flux. Across the water content levels, all the freeze–thaw treatments increased the N2O flux and reduced the CH4 flux as compared to the CK treatment. In addition, long-term freezing reduced the CO2 flux, but the treatment of slowly freezing and quickly thawing increased the CO2 flux. In summary, these results suggest that the freeze–thaw treatments had quite different effects on N2O, CH4, and CO2 fluxes, and their effects on GHG fluxes are more significant than those of soil water content on the eastern edge of the QTP.