Shake table tests are performed on temporary internal partitions for office buildings. Four different specimens are tested. A steel frame is designed to exhibit relative displacements, which typically occur at a given story of ordinary buildings. Four different partition walls are tested simultaneously for each specimen typology. This allows investigating the influence of an innovative device on the seismic performance of the tested components. The innovative device aims at avoiding the unhooking of the panels from the supporting studs. The input to the shaking table consists of two 30-second time histories representative of a target ground motion and acting simultaneously along the two horizontal directions; the time histories are artificially defined so as their response spectra, i.e. test response spectra (TRS), match the normalized required response spectrum (RRS) provided by the AC156 code “Acceptance criteria for seismic qualification testing of non-structural components”. This represents a key issue of the study, involving choices in the standardization of the shake table test procedure on nonstructural components, with particular reference to partitions. Several shake table tests are performed subjecting the specimens to interstory drift ratios up to 1.57%. Both the hysteretic curves and the natural frequency trend highlight that the partitions do not contribute to the lateral stiffness of the test setup. The damping ratio increase after the partition walls are installed within the test frame, causing a beneficial effect in the dynamic response. Minor damage state occurs for interstory drift ratio (IDR) in the range 0.41÷0.65 in standard specimens, whereas moderate and major damage states are attained for IDR in the range 0.51÷0.95. Significant increase of collapse IDR is recorded with the introduction of the innovative device, up to IDR larger than 1.45%. It can be therefore concluded that a simple innovative device is defined, which significantly increases the seismic performance of the tested specimen.