Over seven million students in the United States public school system are served by the special education system, including over 700,000 children between the ages of three and five (National Center for Education Statistics, 2021). These students are legally entitled to the educational opportunities and programs offered to their non-disabled peers, but as the computational thinking field develops best practices for early childhood CT education, best practices for early childhood CT special education are not yet developed. This paper will discuss contrasting current practices for including young children with disabilities in CT classrooms and suggest future directions for research. As multiple approaches to special education exist in research and practice, there is equal variation in approaches to computational thinking education for students with disabilities. One approach refers to explicit instruction, a method of instruction derived from behaviorism(Knight, Wright, & DeFreese, 2019). These practices are very precise, teaching specific skills as behaviors and sequences of behaviors, and are commonly used for students with autism and intellectual disabilities. These practices have been used to teach coding, robotics, and computational thinking to students with moderate and severe disabilities in substantially separated special education classroom (Knight, Wright, & DeFreese, 2019; Knight, Wright, Wilson, et al., 2019; Taylor, 2018). Another common approach is Universal Design for Learning, a process for planning and developing curricula in order to accommodate students with a range of disabilities in inclusive settings (Capp, 2017). Where EBP-based curricula are precise and prescribed, UDL promotes the use of multiple approaches to present knowledge, engage students, and assess learning. Research by Maya Israel has examined use of UDL by teachers and mechanisms by which teachers can incorporate UDL practices into their CT classrooms (Israel et al., 2020). As a field, there is a lack of knowledge as to how students with disabilities learn CT but there is reason to believe these processes are different for students with different disabilities. Young children in the special education system have a range of disabilities, including language delays, autism, learning disabilities, and intellectual disabilities (National Center for Education Statistics, 2021). Some disabilities impact math learning, language processing, or working memory, which may impact computational thinking learning, but we do not know enough about learning processes to make this assessment. As students have a variety of impairments and access needs, one approach to CT learning may not equally serve students with all disabilities. For this reason, researchers should aim to understand how students with different disabilities learn CT to develop best practices for the individual needs of students. As future steps, the authors of this paper will be differentially modifying a computational thinking curriculum to serve students with mild and moderate to severe disabilities in both inclusive and substantially separate classrooms, as well as examining outcomes for these students. Future work should examine how CT assessments can be modified and adapted to accurately assess the CT knowledge of students with different disabilities.