The partition wall systems of interest in this study are used in typical office building occupancies and one of their interesting architectural features is that they can be easily re-arranged on the floor area without damaging the interior building finishing. They comprise rigid panels clipped on steel light gage framing, glazed panels and doors that can be combined in several geometric layouts to create a closed work space. The bottom railing of the framework is attached to carpeted floors with adhesive carpet fastening strips while the partition system is left unrestrained at the top, with only non-structural fastening to the suspended ceiling above it for providing sound insulation. The purpose of the shake table tests is to study the load paths and overall performance of this type of moveable office partitions (MOPs) and verify their seismic capacities for use in office buildings in Canada. Two specimens were tested on the shake table at the Structures Laboratory of the École Polytechnique de Montréal, Canada. They both had the basic plan geometry of a C-shape 3m × 4m × 3m and a height of 2.6 m, which is typical of a single work unit area. Shake table inputs are unidirectional and perpendicular to the long side wall (front wall). They were mainly the top floor responses of two Montréal office building models in SAP2000 to the base accelerograms that match the seismic hazard in Montréal and Vancouver, as prescribed by the 2005 National Building Code of Canada. For comparison, three top floor responses to the 1999 Chi-Chi Earthquake in Taiwan were also selected as the input. The shake table test results indicate that with proper joint detailing and provided the installation is of good workmanship, the MOPs can perform well in moderate to large Canadian design earthquakes, with the maximum floor input level of 1.4g. However, they would not likely resist the near fault seismic events such as Chi-Chi 1999, with large peak floor acceleration up to 3.0g and displacement up to 115 mm.