Parental structure analysis (PSA) is a computer program to analyse separate contributions of paternal and maternal parents to postdispersal plant offspring. The program provides joint estimates of maternal, paternal and cross-parental correlations within and among a set of predefined groups of seeds or seedlings, as well as derivative estimates of effective parental numbers. PSA utilizes data sets that distinguish between maternal and paternal contributions to the genotype of each offspring in the sample, but does not require parental samples per se. The approach requires assay of codominant diploid markers from both seed coat (maternally inherited) and seedling/embryo (biparentally inherited) tissues for each offspring. A simulation analysis of PSA's performance shows that it provides fairly accurate parental correlation estimates from affordable sampling effort. PSA should be of interest to plant biologists studying the interplay between dispersal, demography and genetics, as well as plant–animal interactions.