During the initial phase of COVID-19 pandemic, hospital infection (HI) accidents occurred with alarming frequency. Proactively preventing and controlling HI risks is particularly challenging as hospital is a complex and dynamic scene that involves multiple uncertainties and coupled risks. Thus, this paper develops a multimethod fusion model that combines association rule mining with complex networks to systematically analyze coupled HI risks. Firstly, risk factors of HI are obtained and categorized based on real data in hospitals. An improved Apriori algorithm is used to mine coupled relations between risk factors and generate association rules. Then, complex networks of HI risk factors are constructed based on existing rules. Through an in-depth analysis of the networks’ topology and statistics properties, this study reveals that HI risk networks follow a power-law distribution, attesting to the non-random nature of risks. The proposed approach is applied to analyze HI risks during COVID-19 outbreak in a tertiary hospital in Wuhan. Results show that the method can successfully reveal the coupling of risks and identify key risks of HI. Effective responses can be undertaken in advance to prevent identified HI risks, thereby providing valuable guidance for future hospital management.