Objectives　This study examined the effects of checking the "Tabepo Check Sheet," which has 10 food groups, on changes in dietary variety among older adults living in Japanese metropolitan areas.Methods　We used baseline and two-year follow-up data from a community-wide intervention to prevent frailty in Ota City, Tokyo. A total of 8,635 non-disabled residents (4,145 men and 4,490 women) aged 65-84 years who responded to the self-administered questionnaire surveys in 2016 and 2018, were included in the two analyses. In 2018, we asked about their experience with the "Tabepo Check Sheet," which had been distributed among participating communities from three districts in Ota City since July 2017. The "checked group" consists of those who answered "habitually check" or "have checked." Dietary variety was assessed using the Dietary Variety Score (DVS). We performed propensity score matching with a 1:1 ratio (checked group vs. non-checked group) for the following variables: demographic, socioeconomic, physical, medical, and lifestyle variables, and DVS. After propensity score matching, 876 participants from the checked and non-checked groups were selected. We used a two-way ANOVA to examine the changes in DVS over two years. We also compared the rates of DVS ≤3 and ≥7 points in 2018 using multivariate adjusted logistic regression analysis.Results　In 2018, 11.9% of the respondents had checked the "Tabepo Check Sheet." The DVS mean±standard deviation in the checked and non-checked groups were 3.9±2.2 points and 3.9±2.3 points in 2016, respectively, and 4.5±2.4 points and 4.1±2.4 points in 2018, respectively. The DVS showed a significant interaction between checking experience and time (p<0.001). The rates of DVS ≤3 points in 2018 were 35.2% and 43.8% among the checked and non-checked groups, respectively. The multivariate adjusted odds ratio (OR) of DVS ≤3 points was 0.68 (95% confidence interval [95% CI] 0.56-0.83) in the checked group versus the non-checked group. The rates of DVS ≥7 points in 2018 were 21.7% and 16.8% among the checked and non-checked groups, respectively. The multivariate adjusted OR (95% CI) of DVS ≥7 points was 1.40 (1.10-1.78) in the checked versus the non-checked group.Conclusions　These results suggest that checking the "Tabepo Check Sheet" may improve dietary variety. However, the effects of dietary variety improvement may be weakened by the incidence of skipped meals, social isolation, frailty, and lack of social participation.