Automotive lighting presents a challenging and interesting illumination design space. In addition to non-trivial intensity distribution test specifications, automotive lighting is also increasingly important for communicating stylistic details. Light guides, in which light is introduced at one or both ends and then is extracted at roughly orthogonal angles to the path curve, allow a great deal of aesthetic freedom, and are therefore well suited to automotive lighting tasks that need to convey a brand’s image-for instance, daytime running lamps, tail lamps and turn indicators. Light guides can also help minimize the number of sources and parts needed to create a desired light distribution; or may facilitate a more favorable source placement from a packaging point of view, making them even more attractive for these lighting solutions. Optical designs for automotive light guides are challenging and consequently time consuming. This problem is exacerbated by increasing complexity due to new styling demands, higher expectations for perceived uniformity from multiple viewing directions, and the desire to reduce required source power. In this paper, we describe a new software approach to automatically creating and optimizing light guides with prismatic extractors in a CATIA environment . Often the light guide path curve has a complex shape, and the light guide’s extractor surface must be oriented such that it is opposite the desired light direction. The cross section of the light guide is often partially circular, but other shapes may also be used. Additionally, extractors, which can be bumps or holes, need to be created and oriented appropriately on the surface, and key geometrical parameters that are related to optical performance must be easily changed. Our approach leverages the powerful CATIA environment to construct the light guide geometry. Light sources, ray trace simulations, optical material properties, and optical sensors are also added directly into the CATIA model. Furthermore, many types of optical analysis can be performed after Monte Carlo rays are traced. For instance, one can examine the intensity distribution and select specifications against which to measure test points. Luminance camera, ray history and ray file sensors are also available. Using Monte Carlo ray trace results, the light guide prism geometry is automatically optimized to achieve a desired spatially uniform (or deliberately non-uniform) light distribution. At the same time, the angular distribution is optimized by adjusting prism face angles to point light towards defined angular centroid targets. We employ a binning concept that splits the light guide into sections along its length. Prisms in each bin are associated with the light distribution nearby and are adjusted so that the light from the associated bin has a specified relative flux within a cone, as well as a specified centroid pointing direction. In many cases, the source is only on one end of the light guide; however, in some situations, sources at both ends are needed. Other design considerations include fillet radii, prism face curvature, and draft angles. We provide design examples in the paper.