Bridge scour is the loss of soil by erosion due to water flowing around bridge supports. Scour has been the number one cause of bridge failure in the United States with an average annual rate of 22 bridges collapsing or being closed due to severe deformation. The current bridge design code suggests engineers to use the resistance factor of 1 in the foundation design for extreme conditions. This paper aims to modify the resistance factor in the foundation design in the case of pier scour using reliability theory. First, a set of database is used to quantify the statistical parameters associated with the scatter between the predicted and measured pier scour depth and the probability that a deterministically predicted pier scour depth will be exceeded. The database is also used to provide the bias factors in current predictions. In the second part these statistical parameters are used to develop a reliability-based load and resistance factor design for shallow and deep foundations subjected to scour. The goal is to provide a design procedure where the reliability of the foundation is the same with or without scour.