BACKGROUND: Smoking is a major public health problem worldwide. Research has shown that neighbourhood of residence is independently associated with the likelihood of individuals' smoking. However, a fine comprehension of which neighbourhood characteristics are involved and how remains limited. In this study we examine the relative contribution of objective (police-recorded) and subjective (resident-perceived) measures of neighbourhood crime on residents' smoking behaviours. METHODS: Data from 2,418 men and women participating in the 2007/8 sweep of the West of Scotland Twenty-07 Study were analyzed. Smoking status and perceived crime were collected through face-to-face interviews with participants. Police-recorded crime rates were obtained from the Scottish Neighbourhood Statistics website at the datazone scale. Adjusted odds ratios and 95% confidence intervals were estimated for the likelihood of current smoking using logistic regression models. Adjusted mean daily amount smoked and F statistics were calculated using general linear models. Analyses were conducted for all respondents and stratified by sex and age cohort. RESULTS: Compared to individuals living in low crime areas, those residing in an area characterized by high police-recorded crime rates or those perceiving high crime in their neighbourhood were more likely to be current smokers, after controlling for individual characteristics. The association with smoking was somewhat stronger for police-recorded crime than for perceived crime. Associations were only slightly attenuated when adjusting for either the objective or subjective crime measures, suggesting that these indicators may exert an independent influence on the risk of smoking. Stronger effects were observed for women compared to men. Police-recorded crime rates were more strongly related to smoking status among older respondents than among the younger cohort, whereas the strongest effect for perceived crime was observed among younger participants. CONCLUSIONS: Our findings highlight the relevance of paying attention to both objective and perceived neighbourhood crime measures when aiming to prevent smoking.