Serious games are often characterised as being games developed and used for educational or non-entertainment purposes. The academic literature is saturated with definitions and developed games surrounding the concept. Despite this, there is a lack of quality meta-analytical and longitudinal analysis studies associated with use of serious games. One area where serious games have pedagogical potential is in teaching historical subjects and in displaying historical artefacts. Serious games, combined with evolving 3D modelling techniques and Virtual Reality (VR) capabilities can replicate historical events, artefacts, and environments. The creation of a serious game or even virtual museum that informs individuals about the past have the potential to provide a sense of historical realism or immersion. In the context of Covid-19 and the current global impact to the tourist industry, the application of serious games for such purposes has undoubted significance. This paper reviews the academic literature surrounding the concepts of content and assessment integration in relation to developing historical games for pedagogical and knowledge sharing purposes. The research focuses on two points. Firstly, can serious games exhibit a sense of realism and immersion towards teaching and learning history. Secondly, what are the immersive characteristics needed to make a historical game immersive. Based on the literature this paper will provide a framework that identifies best practice towards making historical games immersive from a games design perspective. The paper also presents an example of a serious game intended to educate users about military machinery used in World War two. A display of 3D modelled Roman artefacts is also presented to document how 3D modelling can replicate an interactive and immersive virtual experience. The findings of the paper argue that there is a close interdependency between aesthetical and content design and learning outcomes towards depicting historical realism and immersion. Future empirical work is also discussed in the paper with the aim of evaluating the framework presented.