A matrix is considered full rank if its rank is equal to the smallest dimension of the matrix, which means all its rows or columns are linearly independent. This property is crucial for solving systems of linear equations and ensures that the matrix has a unique solution or a unique least squares approximation if the system is inconsistent.