We aimed to examine whether long-term use of benzodiazepines is associated with an accelerated decline of cognitive performances by using a statistical model specifically adapted to multivariate longitudinal bounded quantitative outcomes. The data came from the "Three-city" study, a French population based study. All the subjects were 65 years old or older at inclusion and had been followed-up for 7 years. The use of benzodiazepines and cognitive functioning were assessed at each examination phase (baseline, 2, 4 and 7 years). Cognitive decline was analyzed using a nonlinear multivariate mixed model with a latent process. This model makes it possible to assess change over time of the latent cognitive process underlying several neuropsychological tests: Mini Mental Status Examination, Isaacs Set test, Benton Visual Retention Test, and Trail Making Test (A and B), and to describe and account for their metrological properties. Analyses were adjusted for age, center, gender, education, socio-professional status, depression, insomnia, high blood pressure, hypercholesterolemia, alcohol, tobacco consumption and physical activity. Nine hundred and sixty nine subjects who reported taking benzodiazepines for 2, 4 or 7 consecutive years were compared to 4226 subjects who were non-benzodiazepine users. Chronic use of benzodiazepine was significantly associated with a lower latent cognitive level (β=-1.79 SE=0.25 p=<0.001), but no association was found between chronic use and an acceleration of cognitive decline, neither on the latent cognitive process (β×time=0.010 SE=0.04 p=0.81), nor on specific neuropsychological tests. Our results suggest that chronic benzodiazepine use is associated with poorer cognitive performance but not with accelerated cognitive decline with age.