Future exoplanet direct imaging missions, such as HabEx and LUVOIR, will select target stars to maximize the number of Earth-like exoplanets that can have their atmospheric compositions characterized. Because one of these missions’ aims is to detect biosignatures, they should also consider the long-term habitability of planets around these stars. We have developed a framework for computing relative biosignature yields for potential target stars, given a model of habitability and biosignature genesis, and planetary occurrence rates. For different model choices, we find that the stellar populations preferred by our metrics vary drastically in terms of stellar masses and ages. The most physically motivated models for biosignature occurrence depend on the duration that a planet has been habitable, which requires precise stellar evolutionary tracks to accurately assess. To obtain the evolutionary track for a star, one requires an accurate stellar model and precise measurements of stellar properties to fit a model to. We analyze the sensitivity of our biosignature yield metrics and other derived stellar properties, such as masses and ages, to stellar model uncertainties and systematic uncertainties in observed stellar properties. We determine the required precision needed to rank target stars according to our long-term habitability metrics and the extent to which obtaining more precise stellar properties decreases the uncertainty in relative biosignature yields.