The effect of highly immersive virtual reality (VR) on learning is mixed, partly due to technical challenges with VR equipment like head-mounted displays (HMDs) and control devices. These can distract learners from the learning environment and its relevant learning material within. This challenge might cause lower virtual presence, higher negative emotions, and extraneous cognitive (working memory) load; all of which have a negative impact on learning outcome. This study investigates whether hands-on training of the interface improves virtual presence, learning emotions, cognitive load and learning outcomes when learning about chemistry in an immersive VR environment. Seventy-six eighth graders participated in a between-subjects design (hands-on training vs. no training). The results show that pupils with hands-on training (n = 38) showed significantly higher virtual presence, lower extraneous cognitive load and higher learning outcomes than those without any initial hands-on training (n = 38). However, no significant differences in pupils’ learning emotions and in intrinsic and germane cognitive load were found. Implications for the design of highly immersive VR environments are discussed from both a cognitive and an affective perspective.