Abstract: This paper argues that the earthquake precursor research is the basis for earthquake prediction. Objects of the precursor research are individual cases. The major limitation on the earthquake precursor study comes from insufficient observation cases. Since high potential of major earthquake occurrence still exists along the southwestern Longmenshan fault zone, now it is due time to establish a dense network with four-gauge borehole strainmeters (FGBS) over the area. FGBS is a Chinese invention and has already been compared to seismometers. In the earthquake precursor research, FGBS has advantage both in theory and in practice. We think that an FGBS site should be built in the site with stress concentration. Successful establishment of an observation site depends on good condition of the borehole location and right installation of the instrument. The key point of observation with an FGBS is self-consistency of the data from the four gauges. It is unnecessary to pursue the deep borehole strain observation and to conduct absolute stress measurements in order to build a site of FGBS.