The authors present a way to solve the problem of assessing various geodetic stations’ temporal stability using data from continuous GNSS observations. The aim of the study is to compare the design features of the points with the parameters of coordinate time series and to determine the classes of motions that can be reliably detected for different station types. The initial data consist of long-term coordinate time series from GNSS stations installed on different kinds of geodetic objects (ground-based, roof-mounted, and those tapped into engineering structures). The analysis was performed using a time-series model that includes a linear trend, seasonal harmonics, possible offsets, and a residual noise component. For the stations, displacement rates, residual scatter of the horizontal and vertical components, as well as amplitudes of the annual one were estimated. It is shown that the type of geodetic monument primarily affects the structure of the residual signal and the robustness of interpreting the coordinate changes. The most stable results were obtained for foundation-based and deeply anchored points, whereas stations installed on roofs and structures are more susceptible to local structural influences