There are several approaches to estimate possible storage capacities for aquifers and traps in sedimentary basins, ranging from static theoretical capacities estimates to more detailed methods involving dynamic modelling. In this paper, we used a modified version of the basin modelling software SEMI [1, 2] which applies a ray tracing technique to migrate CO 2 within a carrier bed below a sealing cap-rock. We present a modelling strategy for a systematic modelling of maximum trap storage capacities and a mapping of possible "safe" injection localities for a storage unit. Two end-member models regarding the influence of faults were tested. The basin modelling results are compared and validated with results obtained from an reservoir simulation software.Simulations were carried out for the Trøndelag Platform, offshore Norway covering an area of ca. 15,000 km 2 . The slightly north-westwards dipping Middle Jurassic Garn Formation (Fm.) is considered as a good candidate for CO 2 storage. It is widely deposited at the Trøndelag Platform, with a thickness around 120 m and shallow buried (<2 km). It is overlain by a thick shalemudstone sequence (the Middle Jurassic Viking Group), and thick Cretaceous shales favouring a low risk for caprock leakage.Two simulation approaches were tested. First, injection in the Garn Fm. over the whole study area were simulated, to get the maximum total trap storage capacity. The modelling showed a storage capacity of 2.0 Gt with no faults and 5.2 Gt using interpreted faults at top Garn Fm. level as input to the simulations. Secondly, simulations were carried out with 38 CO 2 injection sites. From these, the injection sites which caused migration out of the study area (e.g. upward to the rim of the storage unit, with only Quaternary coverage), where removed. Finally, 7 sites with very low probability for migration out of the area were selected. These "safe" injection sites were mainly mapped in the centre of the Trøndelag Platform where only a few faults are mapped.