Meuse/Haute-Marne underground research laboratory is installed in a region rated at very low seismic hazard for obvious safety reasons. No earthquake of magnitude greater than 3.5 has been reported or identified in the area of interest through historical evidence or records. There are nevertheless a few active faults over the 60 km distance range whose induced traveling vibrations can be detected. Known historical earthquakes (primarily in the Vosges mountainous region) showed a magnitude between 3 and 6 (Brulhet, 2005). Even if the seismic waves, as a transient mechanical input, are of very low levels, it is of interest to characterize their amplitudes in the laboratory to find out if they may have some influence on the design of underground installations. Another issue is to know whether or not seismic waves are amplified or attenuated according to depth and the lithology of the geological overburden. It is known that seismic ground-motion is generally stronger at the surface than underground. For shallow underground structures, Power et al (1996) have estimated the ratio between the maximum acceleration and the surface at different depths; they observed that it decreased by 10% every 15 meters. This result cannot be generalized since it depends largely on the geological setting. More recently Aydan et al (2009) compared the acceleration of the earthquake in L'Aquila in Italy (6.3 magnitude) between a surface station and a station located at the Underground Physics Laboratory at Gran Sasso, 1400m depth. In this case, the maximum acceleration of the surface wave was found to be more than 6 times higher than in the deep gallery. This amplification is due mainly to the interface between soil and atmosphere, where the bulk waves are largely reflected or refracted and while surface waves (Love or Rayleigh waves) are dominant. There are also site effects that amplify the wave amplitude especially in the superficial sedimentary layers. Furthermore, for the same wave, the amplitude of ground motion is inversely proportional to the stiffness of the rock strata. Vibration amplitude in a deep underground structure in a hard competent rock is expected to be lower than in a soft loosened rock. To highlight those phenomenons, a monitoring network composed of three-components accelerometric stations is installed in the laboratory with a frequency range of 0.1 Hz-100 Hz. The network layout consists of a surface station, a station -254 m deep located in the auxiliary shaft (in the limestones of the oxfordian superior), a station in the experimentation gallery -445 m deep (in the mudstones of the oxfordian medium) and finally a station at the main level -490 m deep in the host layer of the oxfordian inferior (Figure 1). The largest acceleration recorded was measured during the earthquake of Besançon the 23rd april 2004 (magnitude 5.5 and 150 km distant) on the station of surface with a peak ground acceleration of 0.02 m.s-2. Since this earthquake, only 1 or 2 earthquakes per year are recorded with a peak acceleration exeeding the 0.0005 m.s-2 triggering threshold. Despite the very low rate of detection, their analysis on the different station of the network shows a systematic amplification of their amplitude and energy on the surface station and to a lesser extent at the mean level -490 m. It is the station in the shaft (-254 m) that measures the lowest solicitation. In parallel, a field study to characterize rock damping properties has been undertaken from data coming from both a temporary accelerometric network and three vertical seismic profiles in deep boreholes located in the neighborhoods. As regards the temporary accelerometric network, data analysis was based on the H/V method that compares the spectrum of the vertical component of the seismic signal with the spectrum of the horizontal components (Nakamura, 1989). H/V Measurements of recorded earthquakes have highlighted a likely weak site effect on the surface with a dominant frequency of 2.5 Hz and corresponding to a the 125 m thick Kimmeridgian marl geological layer overlying to the underground laboratory. If the absolute amplitude of this site effect could not be determined from this preliminary study, it could be estimated with the analysis of the recorded waveforms on the regional stations. Variations in the attenuation properties of seismic waves through the rock cover were estimated from the data of 3 vertical seismic profiles. These show a consistent trapping like effect of waves in the layers with low seismic impedance. This result should however take into account the change in the frequency range for a direct transposition of the seismic sources (14 to 140 Hz) to the characteristics of earthquakes (1-7 Hz). This result calls for further investigations and numerical modeling to specifically quantify its influence.