This paper reviews the current state-of-the-art in facility location models for distribution system design, focusing on continuous location models, network location models, mixed-integer programming models, and their applications. The problem of locating facilities and allocating customers is a strategic issue for companies, encompassing core components such as the location of fabrication and assembly plants, warehouses, and retail outlets. The paper discusses the fundamental assumptions, mathematical models, and solution approaches for these models. Continuous location models, characterized by a continuous solution space and distance metrics, aim to minimize the sum of distances between facilities and demand points. Network location models, on the other hand, compute distances as shortest paths in a graph, considering both demand points and potential facility sites. Mixed-integer programming models are used to model various scenarios, including single-stage, multi-stage, uncapacitated, capacitated, and multi-product models. These models can be further extended to handle dynamic and probabilistic aspects, addressing factors that change over time or are subject to uncertainty. The paper also highlights the computational challenges and solution methods for these models, including exact algorithms, heuristic approaches, and Lagrangean relaxation techniques. It concludes by discussing the practical relevance and limitations of dynamic and probabilistic models, emphasizing the need for empirical data and scenario analysis to make informed decisions in facility location planning.This paper reviews the current state-of-the-art in facility location models for distribution system design, focusing on continuous location models, network location models, mixed-integer programming models, and their applications. The problem of locating facilities and allocating customers is a strategic issue for companies, encompassing core components such as the location of fabrication and assembly plants, warehouses, and retail outlets. The paper discusses the fundamental assumptions, mathematical models, and solution approaches for these models. Continuous location models, characterized by a continuous solution space and distance metrics, aim to minimize the sum of distances between facilities and demand points. Network location models, on the other hand, compute distances as shortest paths in a graph, considering both demand points and potential facility sites. Mixed-integer programming models are used to model various scenarios, including single-stage, multi-stage, uncapacitated, capacitated, and multi-product models. These models can be further extended to handle dynamic and probabilistic aspects, addressing factors that change over time or are subject to uncertainty. The paper also highlights the computational challenges and solution methods for these models, including exact algorithms, heuristic approaches, and Lagrangean relaxation techniques. It concludes by discussing the practical relevance and limitations of dynamic and probabilistic models, emphasizing the need for empirical data and scenario analysis to make informed decisions in facility location planning.