To explore the rehabilitation efficacy of Mirror Neuron System Based Therapy(MNST) instrument training combined with the Promotion of Communication Effectiveness Method (PACE) one-on-one speech rehabilitation model on post-acute phase stroke patients with non-fluent aphasia. Methods: Sixty patients with non-fluent aphasia after sub-acute phase stroke treated in our hospital’s department of rehabilitation medicine from March 2022 to June 2023 were randomly divided into PACE group, MNST group, and combined group, with 20 cases in each group. The MNST group received MNST treatment for 30 minutes, the PACE group received PACE treatment for 30 minutes, and the combined group received both MNST and PACE treatments for 15 minutes each; the treatment frequency and duration were once per day, five days per week, for four consecutive weeks. The Western Aphasia Battery (WAB) and Boston Diagnostic Aphasia Examination (BDAE) assessment results before and after four weeks of treatment were collected and compared among all patients. Results: One case dropped out from the PACE group. Before treatment, there were no significant differences in WAB scores or BDAE grading among the three groups (P>0.05). After four weeks of treatment, intra-group comparison showed that spontaneous speech, auditory comprehension, repetition, naming, and AQ in the WAB significantly improved in all three groups compared to pre-treatment (P<0.01); inter-group comparison revealed: (1) Compared to the PACE group, the combined group had better scores in spontaneous speech, repetition, naming, and AQ (all P<0.05); (2) Compared to the MNST group, the combined group had better scores in repetition, naming, and AQ (all P<0.05); (3) The MNST group had higher repetition scores than the PACE group (P<0.05); other inter-group scores showed no significant differences (P>0.05). After four weeks of treatment, the BDAE grading improved in all three groups compared to pre-treatment (P<0.05). The combined group had a better BDAE grading than both the PACE and MNST groups (P< 0.05), while there was no significant difference between the PACE and MNST groups (P>0.05). Conclusion: For patients with non-fluent aphasia after sub-acute phase stroke, MNST machine training is more effective than PACE-based one-on-one manual training in improving repetition ability. The combination of the two treatments has a better effect than either MNST alone or PACE alone.