Introduction Peer victimisation is a prevalent occurrence in childhood and adolescence and can often have long-lasting consequences. Previous research using polygenic scores (PGSs) have revealed various genetic vulnerabilities as predictive of victimisation in childhood. However, findings were based on self-report and may therefore be influenced by varying self-perceptions. Previous investigations also focused on average victimisation across childhood, and thus do not capture variability in polygenic predictability over time. The present study, therefore, aimed to investigate associations between PGSs and victimisation using separate and combined reports from teachers and peers in childhood, as well as self-reports in later adolescence to explore trajectories of victimisation. Methods Data were derived from the Quebec Newborn Twin Study. Participants were assessed for victimisation using self-reports from 7 to 17 years and using teacher ratings and peer nominations between 7 and 10 years (n = 536). Ten PGSs related to mental health, cognitive abilities and physical traits were examined as possible predictors of victimisation using linear regressions and growth curve models. Results Findings revealed that PGSs associated with victimisation are consistent across informants, but to varying extent according to estimated effect sizes. Self-reported victimisation was predicted by PGSs related to mental health, while PGSs related to cognitive and physical traits had larger effect estimates when predicting teacher- and peer-reported victimisation. The PGS for educational attainment was consistently negatively associated with victimisation across informants, producing the largest effect estimates (β = −.104, 95% CI = −.169 to −.039) when predicting a multi-informant measure of victimisation. No PGS predicted changes in victimisation over time. Conclusion While the PGS for educational attainment is a robust predictor of victimisation, many PGSs are differentially associated with victimisation depending on the informant. Such findings highlight the need to pay close attention to the phenotypic assessment of victimisation, and show that using multiple informants can both strengthen and provide unique insight into how associations may occur.