Background The physicians often confuse the early symptoms of Frontotemporal dementia (FTD) with Alzheimer dementia (AD), leading to misdiagnosis. There are some cognitive tests to discriminate between AD and behavioral variant FTD (bvFTD), and the INECO Frontal Screening (IFS) is a promising test for this purpose. Objective To assess the performance of the IFS to differentiate patients with AD from patients with bvFTD, compared with the Frontal Assessment Battery (FAB). Methods A prospective study with 117 patients of our cognitive unit (35 case-patients with AD, 34 case-patients with bvFTD, and 48 control subjects). They were submitted to the following successive phases of evaluation: 1) screening; 2) dementia diagnosis; and 3) dementia sub-type diagnosis. The IFS and FAB were blind and independently applied by one neurologist to all the participants to end of phase 1 (screening), before to the definitive diagnosis establishment. Results bvFTD showed a lower performance than AD patients on the IFS total score (F(1, 66) = 70.10, p < 0.01) and FAB total score (F(1, 66) = 17.91, p < 0.01). The IFS and FAB showed a sensitivity of 94.12% (95%CI = 80.3–99.2) and 82.3% (95%CI = 65.4–93.2), and a specificity of 94.2% (95%CI = 80.8–99.3) and 48.5% (95%CI = 31.3, 66.1), respectively. The IFS showed significantly superior discriminatory accuracy than the FAB (AuCIFS = 0.98; AuCFAB = 0.73, p < 0.00001). Conclusion The IFS is useful for discriminating between AD and bvFTD patients. The performance of the IFS to differentiate patients with AD from patients with bvFTD is greater than FAB.