To investigate improvement of cognitive domain function in patients with mild cognitive impairment (MCI) by a new cognitive training method. Methods: Total 102 MCI patients with a MMSE score > 24, MoCA score ≤25, and clinical dementia scale of 0.5. Patients were randomly divided into the training group and control group with 51 patients in each. Both groups were given health education and controlled for dementia risk factors. Additionally, the training group underwent systematic training through a new cognitive training method for 3 months. Patients were evaluated at baseline (pre-training), visit 1 (3 months after training), and visit 2 (6 months after training) using the MoCA and MoCA Cognitive Domain Index Score (CDIS) scores. Results: The MoCA, MoCA MIS, MoCA EIS, MoCA VIS, MoCA LIS, and MoCA AIS scores in the training group at different timepoints were significantly different (P<0.05). There were significant differences in the MoCA scores at different timepoints in the control group (P<0.05). In the training group, the MoCA, MoCA VIS, MoCA LIS, and MoCA AIS scores at visit 1 were statistically different from those at baseline (P<0.05); the MoCA, MoCA EIS, MoCA VIS, MoCA LIS, and MoCA AIS scores at visit 2 were statistically different from those at baseline (P<0.05); the MoCA, MoCA MIS, and MoCA EIS scores at visit 2 were statistically different from those at visit 1 (P<0.05). In the control group, only the MoCA scale scores were significantly different at visit 2 compared to baseline (P<0.05). There was a significant difference in the MoCA and MoCA EIS scores between the training group and control group at each timepoint (P<0.05). Conclusion: This new cognitive training method can significantly improve the overall cognitive function and that of each cognitive domain in MCI patients.