The use of neural networks (NNs) has increased tremendously in several areas of engineering over the last three decades. This paper is intended to provide a state-of-the-art survey of NN applications in pavement engineering over the last three decades. The reported studies are briefly summarized under eight different categories: (1) prediction of pavement condition and performance, (2) pavement management and maintenance strategies, (3) pavement distress forecasting, (4) structural evaluation of pavement systems, (5) pavement image analysis and classification, (6) pavement materials modeling, and (7) other miscellaneous transportation infrastructure applications. To maintain consistency, the review was primarily based on archival journal publications although novel application-oriented NN implementations published in peer-reviewed conference proceedings and edited books were also considered. Recent publications focusing on the development and use of hybrid neural systems in pavement engineering were also included in the review. The increasing number of publications in this area of research in combination with other soft computing techniques every year definitely indicates that more and more students, researchers, and practitioners are interested in exploring the use of NNs in the study of pavement engineering problems.