Maximin distance designs were proposed and studied by Johnson, Moore, and Ylvisaker (1990). A maximin distance design is a design in that the minimum distance between any two points is maximized. Fractional factorial designs are commonly chosen according to the generalized minimum aberration (GMA) criterion. The maximin distance and GMA criteria are connected through the generalized MacWilliams identities. We use maximin distance as a criterion to select fractional factorial designs. We study maximin designs and GMA designs for 16, 20 and 32 runs and find that they coincide in many cases. When they are different, we compare them in terms of other criteria such as hidden pro jection property and estimation capacity.