Recent advances in molecular methods and statistical analyses provide weed scientists with powerful tools for examining the genetic structure of weedy plant populations. The value of these studies depends on effective sampling protocols; however, there is little consensus on how to sample plant populations for genetic diversity analyses. In this review, we draw on published literature that incorporates sampling theory and spatial statistics in population genetic analyses to identify key factors to consider when designing a sampling strategy. We discuss how sampling design is affected by research objectives, biology of the study species, population structure, marker choice, and the genetic parameters to be investigated, and we offer suggestions on defining sampling units and developing sampling protocols.