Objective: Primary psychiatric disorders (PsD) can present with symptomatology identical to that of behavioral variant frontotemporal dementia (bvFTD). To date, clinical guidelines do not provide a solution for this diagnostic challenge. The aim of our study was to prospectively determine which demographic, clinical, neuropsychological, neuroimaging, and cerebrospinal fluid biomarkers are important in distinguishing PsD from bvFTD. Methods: Patients with late-onset behavioral disturbances (aged 45–75 years, 73% male) were included based on their scores on the Frontal Behavioral Inventory and the Stereotypy Rating Inventory and followed for 2 years from April 2011 to June 2015. Odds ratios (ORs) were calculated with backward stepwise logistic regression analyses to investigate the association between baseline clinical and demographic variables and the 2-year follow-up diagnosis of PsD (n = 46) (DSM-IV) versus probable/definite bvFTD (n = 27) (International Behavioral Variant FTD Criteria Consortium criteria). We separately measured the association between additional investigations and the 2-year followup diagnosis. Finally, we combined the selected variables to measure the predictive value of both clinical and additional investigations in a single model. Results: Male gender (OR = 5.9; 95% CI, 1.3–26.0), less stereotypy (OR = 0.08; 95% CI, 0.02–0.34), and more depressive symptoms (OR = 1.13; 95% CI, 1.04–1.24) explained 49% of the variance predicting PsD versus bvFTD (χ2 3 = 29.4, P< .001) and correctly classified 82.1% of the cases. Neuroimaging (OR = 0.02; 95% CI, 0.002–0.123) explained 55% of the variance (χ2 1 = 37.5, P< .001) and, in combination with clinical variables, 66.1% of the variance (χ2 3 = 44.06, P< .001). Conclusions: The present study demonstrated that PsD can be distinguished from probable/definite bvFTD with a thorough clinical evaluation by a psychiatrist and neurologist along with use of validated questionnaires for depression and stereotypy; these measures are even more effective in combination with neuroimaging.