© 2019 Copyright held by the owner/author(s). Some recent smartphone apps detect and predict dangerous situations and warn users. Since these apps detect users’ situations by leveraging machine learning methods based on smartphone sensor data, a large amount of training data is required to construct detection models. However, it is difficult to collect such data, especially in dangerous situations. This research proposes a Virtual Reality (VR) application framework for collecting sensor data in virtual dangerous situations in VR space.