Various studies report discordant results regarding the efficacy, parameters, and underlying mechanisms of repetitive transcranial magnetic stimulation (rTMS) combined with cognitive training (CT) on Alzheimer's disease (AD). The objective of the study was to assess the effect of rTMS-CT on cognition, the activities of daily life, neuropsychiatric behavioral symptoms, and metabolite levels beneath the stimulated areas of the brain in patients with AD and to investigate the correlation of metabolic changes (measured with proton magnetic resonance spectroscopy [H-1-MRS]) with clinical outcomes after treatment. Thirty consecutive patients with mild or moderate AD were enrolled and randomly divided into one of the two intervention groups: (1) real rTMS with CT (i.e., real group) and (2) sham rTMS with CT (i.e., sham group). 10Hz rTMS was used to stimulate the left dorsolateral prefrontal cortex (DLPFC) and then to stimulate the left lateral temporal lobe (LTL) for 20min each day for 4weeks. Each patient underwent neuropsychological assessment at baseline (T0), immediately after treatment (T1), and 4weeks after treatment (T2). The ratios of N-acetylaspartate/creatine (NAA/Cr), myoinositol/creatine (mI/Cr), and choline/creatine (Cho/Cr) in the stimulated cortex were measured using H-1-MRS at T0 and T1. Twenty-eight patients were treated with rTMS-CT for 4weeks. Two patients in the sham group withdrew after being treated several times. Compared with the sham group, the cognitive function and behavior in the real rTMS group improved significantly at T1 and T2. In the real group, compared with the sham group, the NAA/Cr ratio in the left DLPFC was significantly elevated (p=0.045); however, in the left LTL, it only showed a tendency toward increase (p=0.162). The change in the NAA/Cr ratio in the left DLPFC was negatively correlated with the change in the cognitive scales of the Alzheimer's Disease Assessment Scale (ADAS-cog). This study indicated a possible modest effect of rTMS-CT on preventing clinical and neuronal functional deterioration in the left DLPFC of patients with AD. The left DLPFC is a better candidate area than the left LTL.