As　educational　reform　efforts　continue,　a　challenge　for　most　adaptive　learning　software　is　how　to　integrate　theoretical　knowledge　with　practice,　especially　in　the　field　of　engineering　education.　Developing　an　understanding　of　how　to　incorporate　virtual　reality　in　education　to　enhance　the　quality　of　learning　has　become　a　research　hotspot　in　the　e-learning　field.　For　this　purpose,　a　novel　adaptive　virtual　reality　learning　method　based　on　learning　styles　model　is　proposed　in　this　paper,　then　a　virtual　reality　learning　system　for　engineering　education　is　constructed　based　on　the　proposed　approach.　In　this　method,　learning　style　are　categorized　via　learning　style　index　to　predict　learning　preferences,　then　differentiated　virtual　learning　environment　of　engineering　education　are　provided　to　students　according　to　their　individual　learning　preferences.　During　the　learning　process,　various　types　of　learning　data　are　recorded　and　analyzed　automatically.　Then　the　prejudged　learning　preferences　are　continually　adjusted　by　the　learning　data　when　the　immersive　learning　experience　is　offered　to　students　by　the　customization　of　learning　scene　and　teaching　material　in　virtual　environment.　To　adapt　to　differentiated　individuals,　the　structure　and　content　of　virtual　learning　environment　are　gradually　optimized,　with　the　ultimate　goal　of　equipping　the　environment　to　meet　an　unlimited　array　of　actual　student　learning　needs.　Compared　with　the　traditional　adaptive　learning　method　based　on　learning　style,　a　comparison　of　experimental　results　shows　that　the　proposed　method　is　more　effective　in　stimulating　learner　enthusiasm　and　can　greatly　improve　student　academic　performance.　©　2019,　Springer　Nature　Switzerland　AG.