This microgenetic study investigated similarities and differences in use and discovery of addition strategies in children with and without mild mental retardation across 24 sessions. Nine children with mild mental retardation in third through fifth grade classrooms and 14 children without mental retardation in kindergarten classrooms were tested individually over 12 weeks (two sessions per week with 12 addition problems per session) and were given no strategy instruction. Overall, children with and without mental retardation showed strategy change across session, progressing from less to more sophisticated strategies and did not differ in the range of strategies, with from one to six different strategies used in both groups. Pretests measures of conceptual understanding of number, including highest number counted to and magnitude estimation problems were the best predictors of accuracy during testing sessions. These results have important implications for educational practices for children with disabilities. (Contains 3 tables and 1 figure.)