Abstract: Full scale models of cold-formed thin-walled steel-profiled steel sheet floors and cold-formed thin-walled steel-gypsum based self-leveling mortar composite floors were subjected to vibration test under dynamic loading including walking and hammer impact, and to static test under 1 kN concentrated load. The effects on the fundamental frequency, damping ratio and mid-span vertical deflection of the composite floors were studied when floor panels were different and steel meshes were installed. The study shows that casting gypsum based self-leveling mortar on the profiled steel sheet could reduce the fundamental frequency, damping ratio and mid-span vertical deflection of composite floors. Nevertheless, it did not significantly increase the dynamic characteristics and decrease the vertical deflection of composite floor via installing steel mesh into gypsum based self-leveling mortar. ABAQUS finite element software was used to conduct modal analysis of the test models as well as variable parametric analysis based on the calibrated model. The research shows that the floor fundamental frequency could be improved and floor mid-span deflection could be reduced by increasing the web height, slab thickness of the floor joist and slab thickness of gypsum based self-leveling mortar, along with strengthening the floor end constraints. In theoretical calculation, the floor could be equivalent to a simply supported beam with uniform mass density and stiffness for predicting the fundamental frequency of cold-formed thin-walled steel composite floors. Additionally, the timber floor deflection calculation formula of Canada was recommended to predict the mid-span deflection of cold-formed thin-walled steel composite floors under 1 kN concentrated load.