Abstract Many regions of southern Australia exposed to broad-acre grazing are denuded of trees. Re-establishing trees in these landscapes is important for both environmental and economic reasons. We investigated whether climate at the site of origin of the seed can inform species selection for tree establishment on ex-pasture sites within agricultural landscapes. We established trials at four sites in the sheep and cattle grazing region of the Midlands of Tasmania at: ‘Woodland Park’; ‘Sorrel Springs’; ‘Inverel’ and ‘Glen Morey’ farms (long-term average rainfalls of 584, 520, 520, and 479 mm, respectively). Eucalypt species were selected, based on high tolerance to frost and drought inferred from climate at the sites of origin and based on the limited scientific literature available, from across Tasmania and mainland Australia. We investigated the mortality, health and height growth of 18 eucalypt species three months, one and six years after planting. Climatic data from the site of origin of the seed of these species was used to group species into categories of mean annual temperature and mean annual rainfall at seed-source. Sensitivity of species to environment was investigated and compared using modified joint regression. At three planting sites, Woodland Park, Sorrel Springs (except health of the temperature category comparison at six years) and Inverel, we found significantly greater height and health one and six years after planting for local (Tasmanian) compared to non-local (Australian mainland) species, and for species where the site of origin of seed was categorised as low or medium mean annual temperature compared with those categorised as high mean annual temperature. There were no significant differences in mortality for these comparisons. At Glen Morey, height was significantly greater at six years after planting of local compared to non-local species and of species from low and medium mean annual temperature categories compared with the high mean annual temperature category. Mortality was significantly greater in local than non-local species six years after planting. Health was not significantly different for these comparisons. Differences in height, health or mortality of species categorised into mean annual rainfall at seed-source were largely non-significant at all sites one and six years after planting. A study of the sensitivity of species to environment of planting indicated that Tasmanian species were less sensitive than mainland species when height data was compared. When health data was compared, species with a site of origin of seed categorised as low rainfall were less sensitive to environment than species categorised as medium and high-mean rainfall. When mortality data was used significant differences were identified in the sensitivity of species to planting environment but this was not related to locality, mean annual temperature or rainfall at seed-source. Species that attained above average height and low mortality across three or more sites included Eucalyptus pulchella, Eucalyptus perineana (Tasmanian species), and Eucalyptus microcarpa, Eucalyptus benthamii and Eucalyptus melliodora (mainland species). Whilst temperature (and to a lesser extent rainfall) at seed-source is a good indicator of the success of tree establishment, high variability between sites points to the need to consider climate, browsing pressure and farm management at the paddock-scale when planning revegetation plantings. This has important implications for carbon sequestration plantings.