School bullying is a significant and widespread problem among school-age children around the world. Because of its psychological and social maladjustment problems, bullying has been gaining the attention of both researchers and the public. In recent years, with the high-speed development of information technology and network technology, a new form of bullying has emerged, that is, cyberbullying. Cyberbullying is a contemporary variant of the traditional bullying, in which individuals intentionally and repeatedly hurt others in an electronic context. In the current studies of cyberbullying, most of them were based on cross-sectional study design. However, longitudinal studies are able to examine the dynamic processes of cyberbullying, as well as risk and protective processes associated with cyberbullying. These longitudinal research conclusions could contribute to the prevention and intervention of cyberbullying. Therefore, the aim of present research is to investigate the potential developmental trajectories and longitudinal risk factors of cyberbullying in adolescence. The present research examined the developmental trajectories of cyberbullying by employing a longitudinal, four-wave design. Moreover, the current research also examined the impacts of moral disengagement, perspective taking and empathy. The sample consisted of 496 adolescents (52.8% boys) in grade eight, who was followed up for two years from grade eight to grade nine. They were assessed on moral disengagement, perspective taking and empathy at Time 1, cyberbullying from Time 1 to Time 4. Using latent growth modeling, the current research examined: (1) adolescents’ initial levels and growth rate in cyberbullying over time; (2) the predictive effects of moral disengagement, perspective taking and empathy on both initial levels and growth rate in cyberbullying. The results of unconditional latent growth modeling analyses indicated that the junior high school students’s cyberbullying presents a decreasing trend significantly in the period of 2 years. Furthermore, the initial levels of adolescents’ cyberbullying were related to the growth rate in cyberbullying. Conditional latent growth modeling was examined with gender, moral disengagement, perspective taking and empathy as time-invariant risk fac--tors on both initial levels and growth rate in adolescents’ cyberbullying. The results showed that (1) moral disengagement has a significantly positive prediction on the initial levels of cyberbullying and negative prediction on the growth rate; (2) perspective taking has a negative prediction on the initial levels of cyberbullying and significantly positive prediction on the growth rate; (3) empathy has no prediction on the initial levels of cyberbullying and the growth rate. The present findings revealed developmental trajectories of adolescents’ cyberbullying. The results also suggested that moral disengagement and perspective taking could play an important role in the development of cyberbullying among adolescence. Overall, the findings support the theory of Moral Disengagement and Social Information-Processing Model.