This causal-comparative study evaluated a 30-week chess instructional program implemented within special education math classes for students in the sixth, seventh, and eighth grades in a suburban middle school located in the southwestern United States. An analysis of covariance (ANCOVA) was utilized to compare the adjusted means for the comparison and treatment groups on the students' math achievement as measured by end-of-year course grades and state assessment scores, the Texas Assessment of Knowledge and Skills (TAKS). Pretest scores and grade levels served as covariates. Results indicated a significant difference on four of the measures in favor of the treatment group: end-of-year course grades, overall TAKS math scale scores, and percentage scores on two specific TAKS math objectives: Numbers, Operations, and Quantitative Reasoning and Probability and Statistics. No significant differences were found between the groups on the other four TAKS math objectives: Patterns, Relationships, and Algebraic Reasoning, Geometry and Spatial Reasoning, Concepts and Uses of Measurement, and Underlying Processes and Mathematical Tools. Causation and generalizability are difficult due to the narrow scope of this study. However, these results are encouraging and suggest chess is a potentially effective instructional tool for students who receive special education services in math. (Contains 3 tables.)