Human activities increasingly fragment natural habitats, placing many species at risk of population decline. This creates an urgent need to preserve biodiversity and maintain ecological connectivity through wildlife corridors. We present two spatial optimization models for corridor design that explicitly incorporate corridor width as a key ecological criterion. The first model minimizes total corridor cost while meeting minimum width requirements, and the second maximizes corridor width under a fixed budget. Together, these models offer a flexible framework for connecting multiple fragmented areas, accommodating both regular and irregularly shaped candidate patches. The minimum-cost formulation solves in reasonable time for large-scale instances. To improve computational efficiency for the more challenging width maximization problem, we develop a binary search-based algorithm and introduce a cut separation technique that integrates spatial information into the mathematical formulation to generate strong initial bounds on achievable corridor width. We evaluate the models and solution methods on a range of computer-generated and real-world landscapes, including an instance focused on maintaining habitat connectivity for the Florida panther.