Epidemiologic data have demonstrated associations of sleep-onset insomnia with a variety of diseases, including depression, dementia, diabetes and cardiovascular diseases. Sleep initiation is controlled by the suprachiasmatic nucleus of the hypothalamus and endogenous melatonin, both of which are influenced by environmental light. Exposure to evening light is hypothesized to cause circadian phase delay and melatonin suppression before bedtime, resulting in circadian misalignment and sleep-onset insomnia; however, whether exposure to evening light disturbs sleep initiation in home settings remains unclear. In this longitudinal analysis of 192 elderly individuals (mean age: 69.9 years), we measured evening light exposure and sleep-onset latency for 4 days using a wrist actigraph incorporating a light meter and an accelerometer. Mixed-effect linear regression analysis for repeated measurements was used to evaluate the effect of evening light exposure on subsequent sleep-onset latency. The median intensity of evening light exposure and the median sleep-onset latency were 27.3 lux (interquartile range, 17.9-43.4) and 17 min (interquartile range, 7-33), respectively. Univariate models showed significant associations between sleep-onset latency and age, gender, daytime physical activity, in-bed time, day length and average intensity of evening and nighttime light exposures. In a multivariate model, log-transformed average intensity of evening light exposure was significantly associated with log-transformed sleep-onset latency independent of the former potential confounding factors (regression coefficient, 0.133; 95% CI, 0.020-0.247; p = 0.021). Day length and nighttime light exposure were also significantly associated with log-transformed sleep-onset latency (p = 0.001 and p < 0.001, respectively). In conclusion, exposure to evening light in home setting prolongs subsequent sleep-onset latency in the elderly.