This study examined the relationship among plasma homocysteine, folate, and vitamin $B_{12}$ levels and neurocognitive function in 118 community-dwelling elderly subjects (mean age, $75.1{\pm}6.7$ years). The Mini-Mental State Examination (MMSE-KC) was used to screen and assess neurocognitive function in the participants. Dietary intake data including the use of dietary supplements were obtained using the 24-hour recall method by well-trained interviewers. Plasma folate and vitamin $B_{12}$ concentrations were analyzed by radioimmunoassay, and homocysteine was assessed by a high performance liquid chromatography-fluorescence method. The proportions of participants with suboptimal levels of plasma folate (< 3 ng/mL), vitamin $B_{12}$ (< 221 pmol/mL), and homocysteine (> $15{\mu}mol/L$) were 16.1%, 5.9%, and 21.2%, respectively. A multiple regression analysis showed that plasma homocysteine was negatively associated with plasma folate and vitamin $B_{12}$ levels. The MMSE-KC test scores were significantly associated with plasma homocysteine and folate, but not with vitamin $B_{12}$, after adjusting for age, gender, body mass index, living with spouse, education, current smoking, energy intake, and chronic diseases such as hypertension, diabetes, thyroid disease, dyslipidemia, stroke, and cardiovascular disease. A general linear model adjusted for covariates revealed that MMSE-KC test scores increased from the lowest to the highest quartiles of vitamin $B_1$, vitamin $B_2$, vitamin $B_6$, vitamin $B_{12}$, and vitamin C intake (p for trend = 0.012, 0.039, 0.014, 0.046, 0.026, respectively). These results indicate that the problem of folate inadequacy and hyperhomocysteinemia are highly prevalent among community-dwelling elderly people and that dietary intake of the B vitamins and vitamin C is positively associated with cognitive function scores.