Masonry screen walls are architectural elements made from bricks that create different patterns to provide shade and natural ventilation to buildings. This paper presents a digital framework for the design of masonry screen walls, whose performance as environmental control elements is optimized while complying with vernacular construction rules. The framework employs shape grammars for creating a generative design system based on existing construction rules. This design system is then translated into a parametric model and connected to a simulation engine that calculates daylight metrics and cooling energy loads. A genetic algorithm is then used to find a family of optimized solutions, with visual feedback being provided to facilitate the understanding of trade-offs between such solutions. The proposed framework is tested with a case study, in which design solutions are generated by manipulating selected design variables to find optimal design solutions in terms of environmental performance.