Lane detection is important for the lane departure warning (LDW) for advanced driver assistance systems (ADAS). Several approaches for lane detection were suggested in the past. However, robustness is still an issue, in case the markings are not arranged in a straight line or when they are occluded. This paper presents a robust vision-based lane detection method. The key idea is to applymethods fromthe target tracking domain to identify lanes in the image space; we use an InteractingMultipleModel (IMM) approach to increase robustness. Our method is based on two phases: a preprocessing phase to extract areas that potentially represent markings and a tracking phase to identify the lanes. In the preprocessing phase, we use regions of interest, median filtering, Otsus algorithm, and image erosion. The tracking is performed in spatial dimension and based on the Interacting Multiple Model (IMM) approach to estimate the lane pixel positions in the image. Two models are used: one for straight lines and one for curves. The simulation results show that this method has good robustness under various road scenarios.