BackgroundThe seedling stage is the most vulnerable period of growth and development for annual weeds and an important target for weed management operations. To address this, several weed emergence models have been developed, but none are commercially available. Therefore, this study aims to develop a web application that implements predictive weed emergence models for eight different weed species, utilizing weather data sourced from public weather stations.ResultsLolium rigidum Gaudin presented a mean root mean squared error (RMSE) value of 8.9, achieving an RMSE value below 15 (success rate) in 84.5% of cases. This result may be attributed to the use of a water potential base, set at −0.4 MPa, to evaluate water availability. Centaurea diluta Aiton achieved an RMSE value below 15 in all situations, with an average value of 9.0. This weed showed higher accuracy at southern sites than northern sites. Conversely, Avena sterilis ssp. ludoviciana (Durieu) Gillet & Magne achieved higher precision at northern sites where no dry periods occurred. The newly developed model for Bromus diandrus Roth. achieved an average RMSE value of 7.7 and a 100% success rate. Papaver rhoeas L. and the three Phalaris species exhibited lower accuracy in this study than in previous ones. Nonetheless, the success rates for Papaver rhoeas and Phalaris paradoxa L. were still above 70%.ConclusionModels for C. diluta, B. diandrus, L. rigidum, Papaver rhoeas and Phalaris paradoxa have demonstrated potential for adoption in commercial production, while Phalaris minor and Phalaris brachystachys models require refinement. © 2023 Society of Chemical Industry.