Theoretical and experimental studies agree that Bi1-xSbx (0.07 <= x <= 0.21) is a three-dimensional topological insulator. However, there is still a debate on the corresponding Bi1-xSbx (111) surface band structure. While three spin polarized bands have been claimed experimentally, theoretically, only two surface bands appear, with the third band being attributed to surface imperfections. Here, we address this controversy using angle-resolved photoemission spectroscopy (ARPES) on Bi1-xSbx films. To minimize surface imperfections, we have optimized the sample growth recipe. We have measured the evolution of the surface band structure of Bi1-xSbx with x increasing gradually from x = 0 to x = 0.6. Our ARPES data show better agreement with the theoretical calculations, where the system is topologically nontrivial with two surface bands.