The keys to implementing the child-friendly concept lies in enhancing community livability, particularly by advocating for low-carbon travel. Currently, communities are still insufficiently considering children’s needs, with many activity spaces and facility layouts failing to meet the requirements for low-carbon travel. Using Shuangliu District in Chengdu as a case study, this research employed both random forest and logistic regression models to examine the effects of travel distance and road environment on children’s travel mode choices. The findings reveal that travel distance holds the highest relative importance in influencing children’s selection of neighborhood activity travel modes, accounting for 22.48%. Additionally, a nonlinear relationship and threshold effect exist between travel distance and certain travel modes, with approximately 3 km serving as the critical intersection point for walking, cycling, and car. Furthermore, feature such as road smoothness, the availability of sidewalks, and traffic density significantly enhance the probability of children choosing to walk to child-friendly facilities. This research seeks to bridge the gap by examining how facility distribution influences travel mode choice and identifying key feature that help urban planners design child-friendly, sustainable urban environments.