Social networks constitute the backbone of information transmission in social media platforms, where link prediction will contribute to managing information diffusion and controlling public opinion. Based on the existing studies in the field of link prediction,this paper starts from the theoretical foundation of deci-sion analysis and presents a novel link prediction method by introducing the utility analysis. In order to solve the problem of parameter estimation,this paper further develops a Markov Chain Monte Carlo method with er-ror degrees and demonstrates its correctness. Based on the selected information from five QQ groups,a com-parison between the proposed method and the classic ones is carried out in terms of the prediction accuracy. The results indicate that the proposed method enjoys satisfactory prediction accuracy because the individual be-havior is considered in this method,and particularly the introduced error degrees would benefit the potential model users in making reasonable decisions by weighing the model’s efficiency and accuracy.