A cross-sectional study was carried out during January-June 2004 to assess the status of HIV infection among brothel-based sex workers in Kolkata city, India. Six hundred and twenty-two sex workers, selected from six brothels, were included in the study to assess their HIV status, and 362 sex workers, a subset of the above population, were interviewed to study their risk behaviour and practices. Blood samples were collected from each sex worker for testing HIV by an unlinked anonymous method. The prevalence of HIV infection was 9.6%, but was much higher among younger sex workers aged 20 years or less (27.7%) compared to the older age group (8.4%). This difference was statistically significant, indicating an association of younger ages with HIV infection [p = 0.006 and odds ratio (OR) = 4.18; (95% confidence interval (CI) 1.2-13.8)]. HIV was not associated with duration of sex work, average number of clients entertained per day, condom use, practice of sex during menstruation, pre-coital examination of penis for visible ulcer/discharge, suffering from sexually transmitted infections, and entertaining clients outside the brothel. There was a big gap between the reported and the evaluated condom use by sex workers. The results suggest that there is a need to develop suitable HIV intervention strategies, considering the socioeconomic and cultural aspects of city sex workers, with a provision for continuous monitoring and evaluation.