Abstract Future direct imaging missions similar to the HabEx and LUVOIR mission concepts aim to catalog and characterize Earth-mass analogs around nearby stars. The exoplanet yield of these missions will be dependent on the frequency of Earth-like planets, and potentially the a priori knowledge of which stars specifically host suitable planetary systems. Ground- or space-based radial velocity surveys can potentially perform the pre-selection of targets and assist in the optimization of observation times, as opposed to an uninformed direct imaging survey. In this paper, we present our framework for simulating future radial velocity surveys of nearby stars in support of direct imaging missions. We generate lists of exposure times, observation time-series, and radial velocity time-series given a direct imaging target list. We generate simulated surveys for a proposed set of telescopes and precise radial velocity spectrographs spanning a set of plausible global-network architectures that may be considered for next-generation extremely precise radial velocity surveys. We also develop figures of merit for observation frequency and planet detection sensitivity, and compare these across architectures. From these, we draw conclusions, given our stated assumptions and caveats, to optimize the yield of future radial velocity surveys supporting direct imaging missions. We find that all of our considered surveys obtain sufficient numbers of precise observations to meet the minimum theoretical white noise detection sensitivity for Earth-mass habitable-zone planets. While our detection rates and mass-sensitivity are optimistic, we have margin to explore systematic effects due to stellar activity and correlated noise in future work.