Cyberbullying remains a critical issue in the digital age, significantly impacting mental health and online safety. This study provides a systematic literature review (SLR) of Scopus-indexed articles from 2021 to 2024, analyzing global trends, roles, and future directions of IoT-based cyberbullying detection. Advanced artificial intelligence (AI) techniques, such as LSTM, CNN, and hybrid models, have proven effective in identifying complex linguistic patterns. The integration of IoT in education presents substantial potential for fostering safer digital environments. However, challenges persist, including adapting detection models to multilingual and multimodal contexts and improving algorithm transparency. Furthermore, addressing social, cultural, and gender dimensions is crucial for developing inclusive and effective solutions. This study underscores the importance of advancing IoT-based models for researchers, supporting educators in promoting digital safety, and guiding policymakers in formulating comprehensive regulations. The findings provide a foundation for future research aimed at enhancing cyberbullying detection systems and ensuring a safer online experience for users across diverse digital platforms.