This work addresses the incorporation of virtual environments as a support to the needs of elementary school students with ADHD. The incorporation of these environments under a multidisciplinary approach can help to strengthen children's skills in a playful and motivating way allowing specialists to work on specific areas of ADHD. For this purpose, a model based on the algorithm for the identification and diagnosis of ADHD is presented from the computational point of view to obtain the design requirements of the proposed virtual environments. A pilot study involving 5 children identified with ADHD and learning disabilities is presented. It was a multidisciplinary team composed of a special education teacher who supports children with learning disabilities and ADHD and the group's teacher in an elementary education institution. Four key ADHD skills (Instruction, Directionality, Memory and Attention) were identified and assessed through the use of the Maze3D virtual environment in which the user used the LeapMotion controller. The results obtained from the evaluation of skills showed an increase in the directionality and memory skills while the instruction and attention skills presented similar or lower values than the pre-test. We also conducted a user experience evaluation which indicated that users were most satisfied with the use of the application.