Planning footsteps before the actual walking motion is a way to ignore the dynamic complexity of bipedal walking in the motion planning process without losing the possibility to exploit the essential characteristic of walking: the fact that it relies on isolated contacts with the ground and therefore provides the ability to traverse irregular or discontinuous terrain. The duality between a continuous environment and discrete sequences of contacts confers to footstep planning a hybrid nature and an interesting theoretical aspect, which adds up to its obvious practical interest. Well incorporated and efficient footstep planning abilities would give humanoid robots a navigation autonomy allowing them to perform a great number of versatile tasks in unstructured environments. Many approaches have been considered to obtain efficient footstep planning, for example, using only a finite number of predefined steps or bounding boxes for fast collision checking. And for really reactive footstep planning, dynamics have to be put back into the equation, either by merging footstep planning with gait control or by adaptively switching between different layers of planning. In this chapter, we give an overview of the main techniques that have been proposed and tested over the past 15 years.