Although numerous studies have been conducted on the responses of soil respiration (<i>Rs</i>) to nitrogen (N) addition in grassland ecosystems, it remains unclear whether a nonlinear relationship between <i>Rs</i> and N addition exists and whether there is a uniform response across grasslands with different degradation status. We established a field experiment with six N treatments (0, 10, 20, 30, 40, and 50 g N m^&minus;2 y^&minus;1) on four grassland sites, each with a varied degradation states in the Inner Mongolia steppe of northern China during the growing seasons of 2012 and 2013. <i>Rs</i> and its major influential factors, including aboveground biomass, root biomass, plant tissues carbon (C) and N concentrations, soil organic carbon (SOC) and soil total nitrogen (STN), microbial biomass and soil pH, were measured. Results show that N fertilization did not change the seasonal patterns of <i>Rs</i> but it changed the magnitude of <i>Rs</i> in grasslands with a different degradation status and only degradation had signification effects on <i>Rs</i>. This shows that variations of <i>Rs</i> in degraded grasslands were due to the difference in SOC content. The response of <i>Rs</i> to N addition differed with the severity of degradation. Furthermore, the response of <i>Rs</i> to N addition slowed down over time. The dominant factor controlling <i>Rs</i> changed across different degradation grasslands. The leading factors for <i>Rs</i> were SOC and STN in non-degraded and moderately degraded grassland; soil pH in severely degraded grassland; and aboveground biomass and root biomass in extremely degraded grassland. Our results highlight the importance of considering the degradation level of grassland to identify soil carbon emissions in grassland ecosystems, and N addition may alter the difference of soil carbon emissions in different degraded grasslands and change its soil carbon emissions pattern.