A model was developed to simulate weed emergence patterns after soil cultivation. In the model, the consecutive processes of dormancy release, germination and pre-emergence growth were modelled in separate modules. Input variables of the model were: date of soil cultivation, soil temperature and soil penetration resistance. Output variables of the model were: seedling density and timing of seedling emergence. The model was parameterized for Polygonum persicaria, Chenopodium album and Spergula arvensis with data from previous field and laboratory experiments. The model was evaluated with data from an experiment, in which emergence of P. persicaria, C. album and S. arvensis was monitored in field plots that were cultivated once only, at one of five dates in the spring. At the same time as the field observations on seedling emergence, seasonal changes in seed dormancy of the buried weed seeds were assessed by testing the germination of seed lots that were buried in envelopes. From a comparison between field observations and simulated data, it appeared that the model overestimated the rate of dormancy release in spring, whereas germination and pre-emergence growth were simulated well. In general, therefore, both the numbers of emerging seedlings and the timing of emergence could be predicted accurately, when dormancy was not simulated but introduced from experimental data. Improvement of predictions of field emergence of weeds should mainly focus on increasing the precision of the simulation of dormancy release. Close correlations were found between seedbed temperature and both the extent and rate of seedling emergence, but analysis with the simulation model revealed that they were only partly based on causal relationships, so that they have limited predictive value.