In this paper, we compare some human face recognition methods which are based l 2 -minimisation description methods. These face recognition methods include the two-phase test sample sparse representation (TPTSSR) method, an improvement to the nearest neighbor classifier and face recognition experiments (INNC), a simple and fast representation-based face recognition method (SFRFR), feature space-based human face image representation and recognition (FSB), collaborative representation classification with regularized least square (CRC_RLS) and relaxed collaborative representation for pattern classification (RCR). We compare all of these methods in the same condition, such as the same database, same test set and same training set. The experimental results show the performance of these six methods.