In ground motion simulation studies based on earthquake rupture scenarios, the absolute ground motion level, frequency content, signal duration and distribution of simulated ground motion is highly dependent on the input parameters used in the calculations. We conduct a predictive study assessing the potential ground motions for Izmir, Turkey based on earthquake scenarios. We calculate ground motions from a reference scenario (Mw6.8) and compare it to 25 test scenarios for the Izmir fault in order to investigate the effect of input parameter uncertainty on the simulated ground motion. In this study, we use a hybrid broadband frequency (0.1–10 Hz) ground motion simulation technique. We find average ground motion levels in the central area of Izmir of more than 600 cm/s2 and 60 cm/s for peak ground acceleration and velocity, respectively, while the standard deviation is found to exceed 200 cm/s2 and 20 cm/s, respectively. The tested parameters identified to have the largest influence on the absolute ground motion level are seismic moment, average stress-drop, rise time and rupture velocity. Parameters identified to have the least effect on ground motion level are rake and ratio of stress-drop on asperities to background stress-drop. The low-frequency ground motion, critical in terms of damage to the building stock in Izmir, is mostly affected by the location of nucleation point, seismic moment, depth of the rupture area, rise time and the velocity model. It is, therefore, highly recommended that future studies focus on reducing the uncertainties in these input parameters.