Young children with severe motor impairments are at great risk of secondary impairments in social, perceptual, and cognitive development. Power wheelchairs can be an effective early intervention tool that provides independent mobility. However, research shows that young children often have difficulties in maneuvering power wheelchairs for everyday use. Existing work on robotic wheelchair assistance has focused on elders and adults with disabilities, who have very different mental and physical characteristics from those of young children. In this project, we will provide robotic assistance for young wheelchair users aged 24 to 60 months. Specifically, we will develop an intelligent shared control framework using state-of-the-art machine learning technology to achieve effective and safe wheelchair maneuverability. This will improve the quality of life for young children with severe motor impairments and reduce the risk of secondary impairments. The project also has great commercial potentials in developing the next generation of intelligent power wheelchairs.