Predicting range expansion of invasive species is one of the key challenges in ecology. We modelled the phenological window for successful larval release and development (WLR) in order to predict poleward expansion of the invasive crab Hemigrapsus sanguineus along the Atlantic coast of North America and north Europe. WLR quantifies the number of opportunities (in days) when larval release leads to a successful completion of the larval phase; WLR depends on the effects of temperature on the duration of larval development and survival. Successful larval development is a necessary requirement for the establishment of self‐persistent local populations. WLR was computed from a mechanistic model, based on in situ temperature time series and a laboratory–calibrated curve predicting duration of larval development from temperature. As a validation step, we checked that model predictions of the time of larval settlement matched observations from the field for our local population (Helgoland, North Sea). We then applied our model to the North American shores because larvae from our European population showed, in the laboratory, similar responses to temperature to those of a North American population. WLR correctly predicted the northern distribution limit in North American shores, where the poleward expansion of H. sanguineus appear to have stalled (as of 2015). For north Europe, where H. sanguineus is a recent invader, WLR predicted ample room for poleward expansion towards NE England and S Norway. We also explored the importance of year‐to‐year variation in temperature for WLR and potential expansion: variations in WLR highlighted the role of heat waves as likely promoters of recruitment subsidising sink populations located at the distribution limits. Overall, phenological windows may be used as a part of a warning system enabling more targeted programs for monitoring.