<b><i>Introduction:</i></b> Ischemic stroke can cause impairment of daily function and cognitive function. Higher cognitive function is reported in many studies to be associated with better functional outcomes; however, evidence from longitudinal study is lacking. Therefore, in the present study, the association between cognitive function and longitudinal changes of functional outcome was investigated based on stroke severity. Furthermore, whether the effect of cognitive function remained consistent after controlling for depression was investigated. <b><i>Methods:</i></b> The data of 423 stroke patients (292 minor strokes, 93 moderate strokes, and 38 severe strokes) were collected. Baseline Mini-Mental State Examination (MMSE) score was considered a predictor, and change of modified Rankin Scale (mRS) score during 12 months of follow-up was the outcome. First, the association between the baseline MMSE score and longitudinal change in the mRS score was analyzed using linear mixed-effects models. Fixed effects were MMSE score group, time, and MMSE score group × time interaction. Additional adjustment was made for the Geriatric Depression Scale (GDS) score. <b><i>Results:</i></b> Among the 423 subjects, the mean age was 73.5 years, and 43.4% were female. In the minor stroke group, the high MMSE score group had a decreased mRS score, and the low MMSE score group had an increased mRS score (<i>p</i> &#x3c; 0.001). This association remained after additional adjustment of the GDS score. Association was not observed between cognitive function and functional recovery in the moderate or severe stroke group. <b><i>Conclusion:</i></b> After ischemic stroke, higher baseline global cognitive function was a predictive factor for better functional recovery regardless of depression symptoms in the minor stroke group.