I present a review of astrometric techniques and instrumentation utilized to search for, detect, and characterize extra-solar planets. First, I briefly summarize the properties of the present-day sample of extrasolar planets, in connection with predictions from theoretical models of planet formation and evolution. Next, the generic approach to planet detection with astrometry is described, with significant discussion of a variety of technical, statistical, and astrophysical issues to be faced by future ground-based as well as space-borne efforts in order to achieve the required degree of measurement precision. After a brief summary of past and present efforts to detect planets via milli-arcsecond astrometry, I then discuss the planet-finding capabilities of future astrometric observatories aiming at micro-arcsecond precision. Lastly, I outline a number experiments that can be conducted by means of high-precision astrometry during the next decade, to illustrate its potential for important contributions to planetary science, in comparison with other indirect and direct methods for the detection and characterization of planetary systems.Subject headings: astrometry -planetary systems -stars: statistics -instrumentation: miscellaneous -methods: statistical -methods: numerical massive, close-in planets. These objects are the easiest to detect with the Doppler method 3 , thus the paucity of high-mass planets on short-period orbits is real, and not due to selection effects.Regardless of the formation mode, orbital migration effects are the likely responsible for the observed M p sin i − P correlation. Many models can reproduce such results, including reduced migration efficiency due to gap opening (Ward 1997;Trilling et al. 2002), substantial mass-loss through Roche lobe overflow (Trilling et al. 1998;Gu et al. 2003), and accelerated orbital decay due to enhanced tidal interactions with the host stars (Pätzold & Rauer 2002). Finally, Ida & Lin (2004a) have derived a theoretical mass-period diagram that closely resembles the one of the extrasolar planet sample, and predicted a paucity of planets in the intermediate mass range 0.05 ≤ M p ≤ 0.5 M J , for orbital distances < 3 AU.The possibility that super-solar metallicity could correspond to a higher likelihood of a given star to harbor a planet has been the subject of a large number of studies (for a detailed review see Gonzalez 2003). Recent works (e.g., Santos et al. 2001Santos et al. , 2004aFischer & Valenti 2005) have conclusively shown that planet occurrence correlates strongly with the host stars' primordial metallicity. Up to ∼ 20% of metal-rich ([Fe/H] 0.3) F-G-K stars harbor planets, while less than 3% of metal-poor stars ([Fe/H] 0.0) have been found to be planet hosts.