Whereas many groups have formal or appointed sub-leaders other than their formal leaders, researchers have focused on the effectiveness of a single leader with a few exception (Kouguchi et al., 2002; Mehra et al., 2006). We examined how leadership behavior by a leader (L) and sub-leader (SL) is effective for the group. Leadership behavior was measured using the framework of PM theory (Misumi, 1984), which measures a leader's degree of task-performance behavior (hereinafter called P) and group maintenance behavior (hereinafter called M). It has been proven that a high P and high M leadership style (i.e., PM style) is most effective. Our first hypothesis is that the complementary leadership style by the leader and sub-leader (i.e., P-L & M-SL, M-L & P-SL), especially the combination of who has more strong social power and M behavior leader with P sub-leader, will be more effective than the opposite one. We investigated nurses and they rated senior (L) and sub-senior nursing officers (SL) leadership effectiveness, on points of PM leadership, social power, inter and intra group leadership and others. As a result of ANOVA, M-L & P-SL, who has strong social power than subleader style was seen to have a high effect on the leadership effectiveness corresponded to perfect style rather than the same leadership style. In addition, complementary leadership style of PM behavior for inter and intra group leadership was effective to leadership evaluation.