Subgradient methods are optimization techniques used for minimizing non-differentiable convex functions by iteratively moving in the direction of a subgradient. These methods extend the concept of gradients to functions that are not smooth, allowing for the handling of a broader class of optimization problems.