This study aimed to investigate the effect of the cognitive conceptual approach as compared to the heuristics approach in developing pupils` problem solving abilities. An experimental study was conducted with two classes of Year 6 pupils from two different primary schools. Both classes were measured based on the improvement they made on their mathematics achievement tests as well as their responses to a survey. A mathematics achievement test containing ten word problems was designed on the classification systems of Riley, Greeno, and Heller (1983) and Carpenter and Moser (1982). Both classes took the achievement test during the first week of school term as a pre-test and during the eighth week of the school term as a post-test. For the survey, the students were administered four subscales of the Fennema-Sherman Mathematics Attitudes Scale (1976) with the aim to measure changes in their attitudes toward mathematics as a result of the intervention. An independent t-test was run as an indicator for statistical significance at the 0.01 level. Pre-implementation results showed that both classes are comparable in terms of mathematics achievement. For the survey, the control group scored significantly higher than the experimental group in terms of self-confidence, motivation, and enjoyment towards the subject. However, at the post-test, a significant difference was found between the two groups with the experimental group reporting significantly higher scores on the mathematics achievement test as well as the four subscales of the survey.