Introduction: Brush cytology is commonly used during endoscopic retrograde cholangiopancreatography for the diagnostic evaluation of biliopancreatic strictures. However, since the overall sensitivity of brush cytology is poor, the exclusion of malignancy is difficult. Recognition of factors related to the patient, technique or lesion may help improve the diagnostic yield of brush cytology. The objective of this study was to evaluate the diagnostic yield of brush cytology in the assessment of biliopancreatic strictures and identify predictive factors associated with a positive diagnosis of malignancy. Methods: Retrospective study that evaluated all consecutive patients that underwent brush cytology for the investigation of biliopancreatic strictures in a tertiary center, between January 2012 and January 2018. Results: One hundred and sixty-five patients that underwent 182 procedures were included. A diagnosis of malignancy was confirmed in 110 patients (66.7%), of whom 62 had positive brush cytology (sensitivity 53.7%, specificity 98.5%, accuracy 69.8%). On the multivariate analysis, age ≥68 years (OR 4.83, 95% CI 1.04–22.37) and lesions suspicious of metastasis on cross-sectional imaging (OR 8.58, 95% CI 1.70–43.38) were independently associated with a positive result. Subanalysis of the patients presenting with these two factors (n = 26) revealed an increase in the diagnostic yield (sensitivity 80.8%). Conclusion: Age ≥68 years and lesions suspicious of metastasis on cross-sectional imaging are independent factors associated with a positive result. Patient selection taking these factors into account may increase the diagnostic yield of brush cytology.