Social responsibility predicts job performance, academic achievement, personality, coping with frustration, self-acceptance and altruism. Previous studies mainly focused on social responsibility’s concept, psychological structure, influencing factors, and instruction. Findings from recent correlational studies suggest that interpersonal relationship influences social responsibility. This study examined the causal relationship between the two variables. Furthermore, prior results postulate that empathy may play an important mediating role in this causal mechanism. Therefore, the present study was designed to test the causal impact of interpersonal relationship on social responsibility and the mediating function of empathy in the association between the two variables. Three studies were conducted. Study 1 was a survey study testing the relationship between interpersonal relationship and social responsibility, and whether empathy played a mediating role, among 335 undergraduates. Study 2 tested the effect of the utility of relationship on social responsibility. 234 undergraduates were randomly assigned to three groups: high utility, low utility and control group. Based on the result of Study 2, Study 3 further manipulated the intimacy of relationship in which participants were asked to imagine a close friend or a classmate who was newly introduced, and examined the effect of utility and intimacy of relationship on social responsibility simultaneously. 192 undergraduates were randomly assigned to four groups classified by the utility (high vs. low) and intimacy (high vs. low) of relationship. The results showed that: (1) Interpersonal relationship correlated positively with social responsibility, and empathy acted as a partial mediating variable. (2) The manipulation of utility of relationship significantly impacted social responsibility. The high utility group showed significantly a higher level of social responsibility than did the low utility and the control group, and the low utility group showed an even lower level of social responsibility than did the control group. (3) The manipulation of intimacy of relationship also impacted social responsibility. In contrast to the low intimacy group, the high intimacy group showed significantly superior level of social responsibility. More importantly, the interaction between utility and intimacy of relationship was significant. Specifically, for the low intimacy group, those with high utility exhibited a significant higher level of social responsibility than did those with low utility. However, in the high intimacy group, no difference in social responsibility was found between the two levels of utility. In conclusion, the present study for the first time confirmed the causal impact of interpersonal relationship on social responsibility and the partial mediating role of empathy. The present results are consistent with the “pattern of difference sequence” account of social responsibility in China.