Four temperature response models were compared describing the emergence rate of Solanum sisymbriifolium (L.) over a broad range of suboptimal temperatures and at different soil water potentials. In the laboratory, the effects were tested on germination rates at constant (9.1–21.8°C) and diurnally fluctuating temperatures at different soil water potentials. Linear, Q10, expolinear and quadratic models were fitted to the data on rate of emergence against temperature. For model validation, field emergence was monitored in 11 sowings conducted in 2001–2004. Emergence rate increased with temperature and was relatively high at soil water potentials in the range of − 0.21 MPa to − 2.6 × 10− 3 MPa, but was almost zero at − 0.96 MPa and − 1.8 × 10− 3 MPa. Diurnal temperature fluctuations did not have a differing influence on germination rates or final germination percentages compared with constant temperatures. The expolinear and the quadratic models were most accurate in explaining variation of laboratory data, especially at temperatures close to the minimum germination temperature of S. sisymbriifolium. These two models had root mean square errors for predicting field emergence rates (5.9 to 38.4 d) of 0.81 and 0.87 d, respectively, and were considered more appropriate to predict the time to 50% germination for crops grown in conditions near their ‘base temperature’ than the widely used linear temperature (thermal time) models. The Gompertz function was fitted to percentage germination versus the time-accumulated germination rate (using the expolinear function to describe the rate–temperature relation). This combined model adequately predicted the temporal pattern of emergence in the field.