The stable nitrogen isotopic composition (δ15N) has been widely used to quantify sources of ammonium (NH4+) in PM2.5. However, the overlap and uncertainty in δ15N values from different NH3 sources, coupled with their seasonal variability, hinder accurate identification of NH4+ source. Here, the δ15N values of various NH3 source samples collected by the active sampler were determined. Subsequently, we measured the δ15N values of NH4+ in PM2.5, which were collected seasonally in Tianjin. We found that the combustion‐related NH3 (c‐NH3) exhibiting higher δ15N values compared to volatile NH3 (v‐NH3), but all δ15N values was fell within the range reported by previous studies. Furthermore, inconsistent seasonal variations were observed in the δ15N‐NH3 values originating from emissions of agricultural soil and human excreta. The application of the Bayesian isotope mixing model (MixSIAR model) revealed a significant increase in the contribution of v‐NH3 to NH4+ when incorporating current source data, as opposed to previous data, for δ15N of NH3 source. Notably, the contribution of v‐NH3 (53.1%) to NH4+ was almost equivalent to that of c‐NH3 (46.9%) when considering the seasonal δ15N signatures of NH3 source. Additionally, the estimated contribution of v‐NH3 to NH4+ exhibited significant seasonal variability, which is more reasonable than in the non‐seasonal scenario. This study demonstrated that v‐NH3 and c‐NH3 contributed to NH4+ in PM2.5 in Tianjin almost equally, and it is highlighted that the seasonal δ15N values of NH3 sources should be considered when estimating the contributions of different NH3 sources to NH4+ in PM2.5 by the MixSIAR model.