The World Health Organization (WHO) claims that 1 in 160 children present an Autism Spectrum Disorder (ASD). However, the real estimation in low- and middle-income countries can be higher and currently is so far unknown. In many Latin American countries, it is essential to develop intelligent tools that can provide support during the psychological diagnosis and intervention of children with ASD. To this aim, this article presents an approach that models the most relevant elements of rehabilitation/educational activities for children with ASD and discuss how to support them through robotic assistants, mobile applications and Multi-Sensory Stimulation Environments (MSEE). The authors tested the proposed system in four different real scenarios: the perception of 47 children without ASD concerning the robotic assistant, the feasibility of using ontology-based reasoning to select therapeutic activities, the opinion of 9 experts on ASD regarding the MSSE and the perception of 36 children with ASD about the robotic assistant.