The aim of this study was to improve the 2nd- and 3rd-graders' conceptual subitizing skills by using in-class activities, and to examine whether the practice affects the students' basic number processing skills, calculation performance and overall mathematics achievement in the long run. The training was carried out for eight weeks (640 min in total). Initially, pretests were administered to assess their basic number processing skills, calculation performance and mathematics achievement. At the end of the training, posttests were implemented so as to analyse their progress. Results showed significant improvements in the experimental groups' basic number processing skills, calculation performance and mathematics achievement. The results of the mathematics achievement test which were administered after a semester to the same groups to evaluate the long-term effects of the training revealed significant progress for both the 2nd- and 3rd-graders. Furthermore, the scores of Tempo-test-rekenen which also administered after a semester to the same group to evaluate calculation performance in the long run improved significantly in the experimental groups. We reached the conclusion that basic number processing skills can be improved with an appropriate intervention that leads to better mathematics learning in the long run.