This investigation aimed to examine the impact of connection stiffness on random vibration testing of heat sink products. The approach involved developing a finite element model of the test piece and fixture, and conducting simulated studies on random vibration testing under varying connection stiffness conditions, in conjunction with actual test results. The findings demonstrated that by adjusting the stiffness of the test piece and fixture models based on real test data, the calculated results closely aligned with experimental outcomes, effectively rectifying inaccuracies in finite element analysis. In summary, it was observed that different forms of connection led to diverse connection stiffness, which emerged as a primary factor influencing overall modal characteristics of both the test piece and fixture as well as significantly impacting simulation outcomes.