Background This study examined the effects of different approaches to teaching learners with mild intellectual disabilities to solve mathematical word problems. Method Thirty Chinese students with mild intellectual disabilities from a special school in Hong Kong were taught using conventional instruction, worked example instruction, and cognitive strategy instruction. A cross-subjects experimental design was used to compare the effects of the three instructional approaches on immediate and delayed tests. Results Students presented with worked example and cognitive strategy instruction solved more problems correctly and generally outperformed students presented with conventional instruction in both immediate and delayed tests. In addition, learners receiving worked example and cognitive strategy instruction were more able to maintain and generalise their previous knowledge and skills to solve novel problems than those receiving conventional instruction. Conclusions The findings indicate that learners can profit from worked example and cognitive strategy instruction which can promote problem-solving skills and mathematical learning.