The present cross-sectional survey of shift-working nurses investigated the characteristics of the Japanese version of the Epworth Sleepiness Scale (JESS), the Stanford Sleepiness Scale (SSS) and actigraphy. Subjects comprised 51 nurses recruited from among participants of the Gunma Nurses’ Health Study. The self-check lists used were the JESS for “recent sleepiness” and the SSS for “current sleepiness.” Actigraphy was performed using an Actiwatch worn for 7–9 days. Subjects also kept a sleep diary. Prevalence of excessive daytime sleepiness was determined using a cut-off value of 11 for total JESS score. SSS was administered at 2-h intervals from the time of starting work until shift completion. Mean (±standard deviation) JESS score was 10.4 ± 3.6; scores were significantly higher for shift workers than for non-shift workers (P > 0.05). Agreement and sensitivity/specificity of JESS score based on maximum SSS score were investigated by varying thresholds for each. JESS score (0–10/11–24) and maximum SSS score (1–3/4–7) showed intermediate agreement (k= 0.416). When the JESS score threshold was 0–10/11–24 and the maximum SSS score threshold was 1–3/4–7, the Youden index (calculated from sensitivity and specificity) was 0.51, indicating those thresholds as cut-off points facilitating detection of sleepiness. Actiwatch-determined sleep variables were uncorrelated to JESS scores, but sleepiness as determined by JESS and actigraphy exhibited a weak correlation. Our findings suggest that the JESS is a scale that reflects SSS, representing current sleepiness, and is an easy-to-use and effective tool for detecting sleepiness. The JESS is useful for self-monitoring, and SSS and Actiwatch can be applied when JESS score is high.