Feature space refers to the multi-dimensional space created by the features used to represent data in machine learning models. Each dimension corresponds to a feature, and the position of data points within this space encapsulates their characteristics, enabling algorithms to identify patterns and make predictions.