Hyposmia is a common finding in Parkinson's disease (PD). The 40-item University of Pennsylvania Smell Identification Test (UPSIT-40) has been adapted and administered in several countries as a diagnostic tool in the diagnosis of PD. We have developed a culturally adapted version of the UPSIT-40 and applied it to 61 nondemented Italian controls and to 68 PD patients. Multiple linear regression analysis was performed to assess the factors that independently influence UPSIT-40 and logistic regression analysis was employed to study the usefulness of UPSIT-40 to predict PD diagnosis. Multiple linear regression analysis showed that PD diagnosis (p < 0.001), age (p = 0.006), gender (p = 0.003) and smoking status (p = 0.03) were significant independent predictors of the UPSIT-40 total score. Using diagnosis as dependent variable, logistic regression analysis showed that UPSIT-40 total score (p < 0.001) was an independent predictor of PD. Using a score ≤ 21/40 as a cut-off point for assigning subjects to PD group, the UPSIT-40 total score differentiated PD and control subjects with 82 % sensitivity and 88.2 % specificity. The adapted version of UPSIT-40 may be useful in addition to clinical examination to improve accuracy of diagnosis of PD in Italian population.