Background: Registered Nurses (RNs) play an important role in caring for patients suffering from cancer pain. A lack of knowledge regarding pain management and the RNs' own perception of cancer pain could act as barriers to effective pain management. Educational interventions that target RNs' knowledge and attitudes have proved promising. However, an intervention consisting of evidence-based practice is a multifaceted process and demands behavioural and cognitive changes to sustain the effects of the intervention. Therefore, our study aimed to investigate if a theory-based educational intervention could change RNs' knowledge and attitudes to cancer pain and pain management, both four and 12 weeks after the start of the intervention. Methods: A quasi-experimental design with non-equivalent control groups was used. The primary outcome was measured using a modified version of the instrument Nurses' Knowledge and Attitudes Survey Regarding Pain (NKAS) at baseline, four weeks and 12 weeks after the start of the intervention to evaluate its persistence. The intervention's educational curriculum was based on the principles of Ajzen's Theory of Planned Behaviour and consisted of interactive learning activities conducted in workshops founded on evidence-based knowledge. The RN's own experiences from cancer pain management were used in the learning process. Results: The theory-based educational intervention aimed at changing RNs knowledge and attitudes regarding cancer pain management measured by primary outcome NKAS resulted in a statistical significant (p<0.05) improvement of total mean score from baseline to four weeks at the intervention ward. Conclusions: The findings of this study, suggest that a theory-based educational intervention focused at RNs can be effective in changing RN's knowledge and attitudes regarding cancer pain management. However, the high number of dropouts between baseline and four weeks needs to be taken into account when evaluating our findings. Finally, this kind of theory-based educational intervention with interactive learning activities has been sparsely researched and needs to be evaluated further in larger projects.