In 1987, a high-resolution aeromagnetic survey was carried out over Lake Superior as a contribution to the Great Lakes International Multidisciplinary Program on Crustal Evolution (GLIMPCE). The survey complemented a 20 s (two-way travel-time) shipborne seismic survey conducted over five profiles on the lake in 1986, the purpose of which was to study the 1.1 Ga Midcontinent Rift under the lake. Two and one-half dimensional (2.5D) magnetic modelling has demonstrated that the model predicted by the seismic results is compatible with the magnetic anomalies under the assumption that the magnetization is due essentially to a normally polarized upper layer and a reversely polarized lower layer. An approach has been developed and used to extend the model based on GLIMPCE seismic data westwards over the central portion of the lake by shifting the 2.5D model laterally to the adjacent north–south profile, adjusting the node points, and repeating the process until the area is covered. A surface integration was then carried out over the resulting three-dimensional (3D) structure. The modelling was constrained primarily by 8.0 s (two-way traveltime) seismic lines provided by industry, which are capable of defining the upper surface of the normally magnetized layer and the boundary between the normally and reversely magnetized layer. Values for the magnetization strengths and directions of both layers were estimated from published values on the shore and on islands in the lake. The model was adjusted to match the measured field as closely as possible, thus providing an estimate of the topography of the lower surface of the reversely magnetized layer. The final model was found to be compatible with the gravity anomaly pattern and seismic refraction studies. Varying the magnetization of the upper and lower layer within reasonable limits was found to change the absolute value of the maximum depth (48 km) of the model by approximately 10% but not to significantly effect the shape. The 3D modelling program was shown to be useful for studies of this type where continuity along strike is reasonable.