A neural network (NN) model is developed to map the design dimensions of complementary split ring resonator (CSRR) printed on a specific substrate with its resonant frequency. This trained NN was then used to find the structure of CSRRs to be etched on the ground plane of a patch antenna to obtain size reduction. Because the response of NN is very fast and as it bypasses the repeated use of the simulator, faster versions of CAD models can be developed for metamaterial based planar antennas.