Soot formation was investigated both numerically and experimentally using a premixed burner-stabilized ethylene–oxygen–argon stagnation flame. Soot modeling is inherently complex and highly sensitive to the choice of the reaction mechanism. This study assesses the impact of reaction mechanisms on the modeling of gas-phase properties and particle size distribution functions (PSDFs) using two widely adopted polycyclic aromatic hydrocarbon (PAH) mechanisms, ABF and KM2. Experimentally measured temperature profiles and PSDFs were compared with numerically predicted results across various burner-to-stagnation plate separations. A clear transition from unimodal PSDFs at shorter separations to bimodal distributions at larger distances was observed experimentally. The numerical results were analyzed to evaluate their ability to predict both the flame temperature profiles and key features of the soot PSDFs. Differences in PAH reaction pathways between the two mechanisms led to distinct particle distribution trends at each separation distance. Both mechanisms captured the transition from unimodal to bimodal PSDFs, but neither fully reproduced the key features of experimental distributions. The KM2 mechanism showed better agreement with the observed trends, particularly at larger burner-to-stagnation plate seperations.