Polishing experiments were carried out on eight different carbonate stones with a machine designed for simulating industrial scale tile polishing process. Polishing was performed at 0.8, 1.3, 1.8 and 2.3 m/min belt speeds under fixed rotational speed and pressure of a polishing head. During the tests, calibrated stone tiles were sequentially polished with 60, 80, 120, 220, 320, 400 and 600 grit SiC abrasive stones at each belt speed. After polishing of each abrasive size, the quality of surfaces was assessed by measuring the surface roughness and gloss. As abrasive grit size gradually reduced to 320 grit, the roughness on the stone surfaces decreased greatly while gloss readings did not change significantly. But after 320 grit, gloss values increased greatly while change in roughness values was not significant. Reducing the belt speed resulted with a significant decrease in roughness for each grit size of abrasive, but after polishing with abrasive size larger than 320 grit, the influence of belt speed on the reduction in roughness values became less. Among the carbonate stones, micritic limestone tiles, which were homogenous in mineral composition with strong and hard structure, resulted with good surface finish. But clay minerals existing in other limestone and heterogeneous porous structure of travertines had an adverse effect on surface quality of tiles. Results from polishing tests suggest that belt speed value of 1.3 m/min provide optimum polishing quality conditions since optimum consistency in roughness and gloss of tile surfaces was reached at this belt speed. This findings were very valuable information for operating factories in which polishing is carried out. The gloss reading on stone surfaces was found to increase exponentially with decrease in surface roughness. Good correlation was found between surface roughness and gloss quality of micritic limestones and marbles compared to travertines. (C) 2010 Elsevier Ltd. All rights reserved.