Background Environmental modification is highlighted as one of the most effective method in child injury prevention, and environmental modification requires analysing and predicting possible injury situations corresponding to not only products themselves but also layout and interaction of them. However, guardians rely not on injury data or children’s development data but on their experience and subjectivity now.

Methods We developed a system to analyse injury situations corresponding to layout of products and furniture. First, the developed system creates statistical model of injury by performing natural language parsing for big data on injury situation collected in medical institutions. Then based on both the created statistical model of injury and children’s development data, the system predicts possible injury situations in a target layout of products

Results The current system allows a user such as a parent and a designer to easily estimate climbable areas or objects that can be reached by a child. The visualisation of climbable areas and reachable objects informs the user about possible injury situation in a target environmental layout. To demonstrate the validity of the developed system, we compared the actual injury data with the prediction. The detailed data on actual injury and situation information were collected by home visit investigations of four houses, and the simulation succeeded in predicting 13 of 14 actual injuries.

Conclusions In this study, we developed injury analysis and prediction system. The system allows a non-specialist to easily grasp a possible serious injury situation corresponding to a target layout of products and furniture. It is useful for determining the arrangement of products and furniture and the storage place of dangerous objects for injury prevention.