All methods have their weaknesses and strengths- there is no one perfect method. None will give you a (large) tree with branches and lengths that have a 100% probability of being correct. Nei's book "Molecular Evolutionary Genetics" has a nice discussion on the appropriate use of various methods.