Horqin Sandy Land is one of the serious desertification areas in north China. Understanding the response of the sandy grassland biomass accumulation to precipitation changes is beneficial to the ecological restoration and ecosystem management for this area. In our study, we simulated the response of vegetation biomass accumulation process to precipitation change in the growing season of Horqin sandy grassland on point scale by combining the vegetation-soil water coupling model and the vegetation Threshold-Delay response model (T-D). The results revealed that (1) the accumulation of vegetation biomass was improved with the increase of precipitation, and otherwise inhibited with the decrease of precipitation. However, under the same degree of change in precipitation, the response of biomass accumulation to the increase of precipitation was far greater than the response to the decrease of precipitation, which revealed that there was a significant nonlinear responding relationship between the accumulation of vegetation biomass and the change in precipitation. (2) The response of biomass accumulation to change in precipitation frequency was significantly positively correlated with the change of single effective precipitation in both dry year and wet year, weakly correlated with the change of cumulative effective precipitation, and significantly correlated with change in effective precipitation interval only in dry years. This result showed that the change of precipitation frequency in different precipitation years actually affects the accumulation of biomass by changing the single effective precipitation and effective precipitation interval. (3) Vegetation biomass accumulation process had obvious response threshold to precipitation frequency change, and this response thresholds were different under different precipitation amount and precipitation characteristics. In all, there was obvious response of biomass accumulation to precipitation change, and the combination of vegetation-soil moisture coupling process model and T-D model can effectively identify this response on daily scale, which provided a new tool for exploring the relationship between vegetation and precipitation.