Objective To explore the efficacy of different forms of repetitive transcranial magnetic stimulation (rTMS) on post-stroke non-fluent aphasia. Methods A total of 45 patients with post-stroke non-fluent aphasia who were hospitalized in Neurovascular Center of The First Affiliated Hospital of Naval Medical University (Second Military Medical University) from Jan. 2021 to Oct. 2022 were randomly assigned (1:1:1) to pseudo-stimulus group, unilateral rTMS treatment group, or bilateral rTMS treatment group. Patients in the pseudo-stimulus group received conventional language training, patients in the unilateral rTMS treatment group received unilateral rTMS therapy on the basis of conventional language training, and patients in the bilateral rTMS treatment group received bilateral rTMS therapy on the basis of conventional language training. The spontaneous speech score, listening comprehension score, naming score, retelling score and the aphasia quotient (AQ) in the Western aphasia battery (WAB) of the 3 groups were compared before and after treatment. Results Before treatment, there were no significant differences in WAB scores (spontaneous speech, listening comprehension, naming, or retelling) or AQ among the 3 groups (all P>0.05). After treatment, the WAB scores (spontaneous speech, listening comprehension, naming, or retelling) and AQ of the 3 groups were significantly increased (all P<0.05); the WAB scores (spontaneous speech, listening comprehension, naming, or retelling) and AQ of the 2 treatment groups were significantly higher than those of the pseudo-stimulus group (all P<0.05), but there were no significant differences in the above indexes between the unilateral or bilateral rTMS treatment groups (all P>0.05). Conclusion Conventional language training combined with rTMS therapy can significantly improve the language function of patients with post-stroke non-fluent aphasia, but the effect of bilateral rTMS is not superior to that of unilateral rTMS.