Urban Transportation Network Design (TND) is a highly complex system engineering. It has always been regarded as a very difficult problem in urban planning and transportation. This study tries to sort out and summarize the existing research system and development process in this field. From the perspective of mathematical modeling, the multi-objective property, nonlinear characteristics, uncertainty and NP-hard computational complexity of the TND problem are explained one by one, and a series of solution difficulties derived from them also appear. Classical optimization methods, such as linear programming, integer programming, bi-level programming models and so on, have advantages in solving structured problems. But their limitations in large-scale dynamic situations are also pointed out. In comparison, new intelligent and simulation optimization methods, such as heuristic algorithms, data-driven modeling strategies and simulation-optimization coupling frameworks, provide more potential ways to solve complex traffic network design problems. The current mainstream research paradigm can be summarized as the progressive process of "theoretical analysis—optimization solution—simulation verification". Different tools work together in this process to deepen the understanding of the problem and improve the solution efficiency. Network planning with dynamic and multiple uncertainties still faces many unsolved difficulties. Big data technology, artificial intelligence methods and the interdisciplinary integration may become the key driving forces for the future breakthrough in the TND field. This review aims to provide basic theoretical reference and framework support for the follow-up related research.