Working memory (WM) deficiency is a primary reason for the poor academic performance of children with learning disabilities (LDs). Studies have shown that the WM of typical children could be improved through training, and WM training contributes to improving their fluid intelligence and academic achievement. However, few studies have investigated WM training for children with LDs, and results have been inconsistent. The present study examined the long-term effects of WM updating training and whether it can mitigate LD symptoms. Fifty-four children with LDs were recruited and divided randomly into a training or control group. The training group underwent adaptive running WM training for 20 days. Before and after training, the 2 groups completed a 2-back task, a digit span task (forward and backward), Raven's "Standard Progressive Matrices" test, and a scholastic attainment test (Chinese and math). The tests were repeated 6 months later. The results showed that, as compared with the controls, the training group exhibited significant improvements in the digit backward span task, 2-back task, and Raven's "Standard Progressive Matrices." The math scores of the training group improved significantly by 6 months after the training. The results of this study suggest that WM updating training could mitigate the cognitive deficits of LDs and improve the WM capacity, fluid intelligence, and math scores of children with LDs. Moreover, the training effects could be maintained for at least 6 months.