Empowering teachers through sharing communal decision-making responsibility via distributed leadership has been shown to be effective for positive change in schools. While studies have proposed various psychosocial channels through which positive effects on teacher wellbeing can be realized, there is scarce evidence on how this relationship is influenced by teacher self-efficacy. This study examines how self-efficacy mediates the relationship between distributed leadership, job and career wellbeing among secondary school teachers, employing a partial least-squares structural equation model using the Teaching and Learning International Survey (TALIS) Shanghai dataset (N = 3799). Results show that distributed leadership is positively associated with improvement in self-efficacy (std. β = 0.33, P < 0.001), job wellbeing (std. β = 0.51, P < 0.001), and career wellbeing (std. β = 0.45, P < 0.001), whereas self-efficacy is positively correlated with job wellbeing (std. β = 0.15, P < 0.001), but not career wellbeing (std. β = −0.01, P = 0.69). In terms of mediation effects, self-efficacy positively mediates the relationship between distributed leadership and job wellbeing (std. β = 0.05, P < 0.001), but distributed leadership does not indirectly influence career wellbeing (std. β = −0.002, P = 0.70) via channels through self-efficacy.