Donepezil has been approved for the treatment for mild-to-moderate Alzheimer's disease (AD), but the therapeutic response rate varies from 20 to 60%. A higher oral dosage was suggested to have a better therapeutic response in reported results, but the plasma concentration of donepezil was not examined with respect to the therapeutic outcomes in those studies. Therefore, we analyzed the therapeutic responses, measured by neuropsychological assessments, among 70 newly diagnosed AD patients taking donepezil (5 mg daily) in relation to their plasma concentration of donepezil, apolipoprotein E genotype, and demographic characteristics. Our results have showed 60% of recruited AD patients improved in cognition, measured by Mini-Mental Status Examination (MMSE), and 57.1% in global status, by Clinical Dementia Rating Scale (CDR) sum of boxes (CDR-SB). In cognition, compared to the improving group, the clinically worsening group had a significantly higher donepezil concentration [p = 0.022, odds ratio (OR) = 1.024, 95% CI = 1.003-1.045] and higher initial MMSE score (p = 0.007, OR = 1.330, 95% CI = 1.080-1.639). In global status, initially higher CDR-SB (p = 0.028, OR = 2.318, 95% CI = 1.096-4.903) and initially higher MMSE (p = 0.036, OR = 1.201, 95% CI = 1.012-1.425), not donepezil concentration (p = 0.883), were significantly associated with clinical worsening. Our results have indicated that the dosage of donepezil should be reconsidered for AD patients, especially those clinically worsening in cognition.