Numerous studies have explored the effects of virtual reality (VR) on adults' cognition. Little is known, however, of these effects in children. The aim of this study was to explore, in both children and adults, the respective roles of the specific factors of VR, such as immersion, sense of presence and emotion, on memory performance. To do so, we used a head-mounted display to present a VR experience in which we manipulated immersion by varying 3D asset quality (High and Low) and emotion by presenting negative, neutral and positive stimuli. 48 adults (M age = 20.65) and 40 children (M age = 11.63) were both divided into two experimental groups (High vs. Low 3D model quality). Valence, arousal, and sense of presence were self-assessed by means of questionnaires, while memory of the presented stimuli was assessed using a free recall task. We also performed physiological measurements to provide objective support for our data. Results showed that memory performance was better for emotional than for neutral stimuli regardless of age group, even though children seemed to avoid looking at negative stimuli compared to neutral ones. Memory was predicted by arousal and presence in adults and only by arousal in children. Memory was not impaired by using poor image quality when highly arousing content was displayed. This study revealed that, contrary to adults, the use of poor image quality did not protect children from strong emotional experiences in VR. The roles of familiarity and arousal are discussed to help explain these results.