Numerous empirical studies have contributed to the understanding of factors connected to students' bystander behaviors in peer victimization situations. Nevertheless, a crucial gap remains concerning the scarcity of longitudinal studies. Drawing on social cognitive theory, the present study examined whether moral disengagement and defender self-efficacy predicted bystander behaviors a year later. Participants were 1346 Swedish adolescents (MageT1 = 13.6 years, MageT2 = 14.6 years) who answered a web-based self-report questionnaire in seventh and eighth grades. Random intercept models showed that higher levels of moral disengagement in seventh grade were associated with more pro-aggressive bystanding in eighth grade (Est = 0.19, p < .001), with interaction analyses revealing that this effect was particularly pronounced in students with high defender self-efficacy (Est = 0.05, p < .01). The results also revealed that higher levels of defender self-efficacy in seventh grade were associated with more defending (Est = 0.18, p < .001) and less passive bystanding (Est = -0.11, p < .001) in eighth grade. Interaction analyses further demonstrated that the negative association between defender self-efficacy and passive bystanding was significant only at low levels of moral disengagement (Est = 0.09, p < .001). Our findings suggest that moral disengagement is more strongly related to the inhibitive form of moral agency among bystanders, whereas defender self-efficacy is more strongly related to proactive moral agency. Thus, interventions aiming to reduce pro-aggressive bystanding and promote defending need to consider both moral disengagement and defender self-efficacy.