Introduction: The lack of longitudinal data of comorbidity burden makes the association between comorbidity and cognitive decline inconclusive. We aimed to measure comorbidity and assess its effects on cognitive decline in mild to moderate dementia. Methods: This was a prospective cohort study. The participants were enrolled from the Hualien Tzu Chi Hospital between January 2015 and December 2018. We enrolled 175 older adults with mild to moderate dementia and conducted in-person interviews to follow-up comorbidity and cognitive function annually. The comorbidity burden indices included Cumulative Illness Rating Scale for Geriatrics (CIRS-G), Charlson Comorbidity Index (CCI), and Medication Regimen Complexity Index (MRCI), and cognitive function was measured by Mini-Mental State Examination (MMSE) and clock drawing test. We employed the generalized estimating equations to assess the longitudinal effect of time-varying comorbidity burden on cognitive decline after adjusting for age, sex, and education. Results: Most patients were diagnosed with Alzheimer’s disease (88.6%) and in the early stage of dementia (Clinical Dementia Rating [CDR] = 0.5, 57.1%; CDR = 1, 36.6%). Multimorbidity was common (median: 3), and the top 3 most common comorbidities were osteoarthritis (67.4%), hypertension (65.7%), and hyperlipidemia (36.6%). The severity index of CIRS-G was significantly associated with cognitive decline in MMSE after adjusting for age, sex, and education. CCI and MRCI scores were, however, not associated with cognitive function. Conclusion: The severity index of CIRS-G outperforms CCI and MRCI in reflecting the longitudinal effect of comorbidity burden on cognitive decline in mild to moderate dementia.