Abstract: Prefabricated buildings are susceptible to multiple interrelated factors with complex interactions during construction, leading to significant uncertainties in key performance indicators (KPIs) such as cost, quality, and schedule. These uncertainties hinder the widespread adoption of prefabricated construction. To address this challenge, this study investigates the formation mechanism of multi-dimensional construction performance uncertainties. Based on critical influencing factors identified through literature analysis, a social network analysis (SNA) model was developed using UCINET to characterize the complexity of the performance uncertainty network. The results reveal a decentralized network structure with low nodal constraint forces, indicating strong autonomous behavioral tendencies among influencing factors. Furthermore, a system dynamics (SD) model was constructed via Vensim and validated through a case study project. Simulation results identify “management competency of personnel” and “effectiveness of personnel management systems” as the most sensitive factors affecting multi-dimensional performance outcomes. The proposed hybrid methodology and findings provide theoretical foundations for formulating targeted strategies to mitigate performance uncertainties in prefabricated building construction.