A template-map design strategy for generating sets of non-interacting DNA oligonucleotides for applications in DNA arrays and biosensors is demonstrated. This strategy is used to create a set of oligonucleotides of size s with length l that possess at least n base mismatches with the complements of all the other members in the set. These "DNA word" sets are denoted as nbml-mers or l:n sets. To regularize the thermodynamic stability of the perfectly matched hybridized DNA duplexes, the l-mers chosen for all the sets are required to have an approximately 50% G/C content. To achieve good discrimination between each DNA word in each set generated using the template-map strategy, it is required that n should be approximately equal to l/2 or higher. The template-map strategy can be used in a straightforward manner to create DNA word sets for cases when l = 4k and n = 2k, where k is an integer. Specific examples of 4k:2k sets are designed: an 8:4 set (s = 224), a 12:6 set (s = 528), a 16:8 set (s = 960), and a 20:10 set (s = 1520). These sets are further optimized to achieve the narrowest possible distribution of melting temperatures by selecting the best set after permutation of the templates and maps over all possible configurations. To demonstrate the viability of this methodology, a non-interacting set of four specific 6bm 12mers have been chosen, synthesized, and used in an SPR imaging measurement of the hybridization adsorption onto a DNA array, The template-map strategy is also applied to generate DNA word sets for cases where l not equal 4k. In these cases, the creation of the maps and templates is more complicated, but possible. The templates and maps for three additional types of sets are created: (4k - 1):(2k - 1), (4k + 1):2k, and (4k - 2):(2k - 1). Specific examples are given for l = 7, 9, and 10: DNA word sets of 7:3 (s = 224), 9:4 (s = 360), and 10:5 (s = 132).