Dementia with Lewy Bodies (DLB) has recently been recognized as the second most common neurodegenerative dementia after Alzheimer's disease (AD). However, studies on costs of care for dementia patients have focused almost exclusively on patients with AD, with limited effort to examine costs of DLB patients. The objectives of this study are to compare total costs of care and its major components for patients with AD or DLB. Data consisted of 170 patients with probable AD, diagnosed according to the NINDS-ADRDA criteria, and 25 patients with DLB, diagnosed according to the 1996 consensus guidelines. At study entry, all patients lived in the community and had a Folstein Mini-Mental State examination (MMSE) score≥16. Patient and informant reported on patient's use of direct medical care (hospitalizations, outpatient tests and procedures, assistive devices, and medications), direct non-medical care (home health aides, respite care, adult day care), and informal care. Patients' clinical and demographic characteristics included global cognitive status (measured by MMSE), functional capacity (measured by Blessed Dementia Rating Scale, BDRS), psychotic symptoms, behavioral problems, depressive symptoms extrapyramidal signs, comorbidities, age, and sex. Costs were compared using covariate matching methods, first matching on age and sex, then additionally on MMSE, and last additionally on BDRS. Average annual costs for caring for DLB patients ($35,143) were higher than that for AD patients ($25,129) but differences were not statistically significant. Direct non-medical costs were significantly lower ($947 vs $1,478) and indirect costs were significantly higher ($23,036 vs $17,136) for DLB patients compared to AD patients. After matching on age and sex, between-group differences for indirect costs and direct non-medical costs remained statistically significant. After additional matching on MMSE, between-group differences for indirect costs were only marginally significant, and for direct non-medical costs they were no longer statistically significant. After additional matching on functional status, between-group differences in all costs were no longer statistically significant. Apparent cost differences were largely attributed to differences in patients' cognitive and functional status. However, small sample size for DLB patients may have limited power to detect statistically significant differences in costs of care between these groups.