The present study examined the mere exposure effect of olfaction, focusing on the relationship between participants' liking for and impression of odorants. Thirty-two college students were exposed to either jasmine essence (Jasmine Group) or rose essence (Rose Group) consistently for 4 successive nights for the entirety of the sleep period. Participants' liking for and impression of both stimuli were evaluated 3 times: before the first exposure, after 2 nights of exposure and after 4 nights of exposure. A significant increase in liking was observed only in the Jasmine Group, in which participants reported low liking at the first evaluation; no effect was seen in the Rose Group, in which participants reported high liking at the first evaluation. As for the impression of the odorants, repeated exposure to rose decreased subjective intensity, and repeated exposure to both jasmine and rose decreased subjective stress. Correlation and multiple-regression analysis revealed that the decrease in subjective stress contributed significantly to the increase in liking. These findings are consistent with the conditioning hypothesis for the mere exposure effect by Zajonc (2001); however, this may be a result of the idiosyncratic properties of olfactory sensation.