This study examined the role of artificial intelligence (AI) in supporting the mathematics education of students with special educational needs (SEN). Thirty peer-reviewed articles were analysed to examine (i) the types of AI technologies implemented in mathematics education, (ii) evidence on the effectiveness of AI interventions in improving learning, (iii) the benefits and limitations of AI in special education, (iv) comparisons of AI applications with traditional methods in terms of personalisation, engagement, and accessibility, and (v) applications of AI in supporting specific disabilities. The findings indicated that intelligent games, chatbots, personalised learning systems, and intelligent tutoring systems were the most frequently utilised AI-powered tools. All studies assessing effectiveness reported positive impacts on mathematical achievement, as well as benefits such as reduced mathematics anxiety, increased motivation, and enhanced independence among learners. Reported challenges included high costs, ethical concerns, teacher preparedness, and a digital divide limiting equitable access. Compared to traditional methods, AI offers more accurate diagnosis of learning difficulties, greater opportunities for personalised learning, and increased accessibility. AI showed promise in addressing specific disabilities such as dyscalculia and autism through adaptive and diagnostic tools. The review concluded that AI should be viewed as a complementary tool that extends teachers’ capacity to individualise mathematics instruction for students with SEN.