A novel Artificial Neural Network (ANN) based two Substrate integrated waveguide (SIW) bandpass filters comprising Complementary Split Ring Resonators (CSRRs) are proposed in this paper. These CSRRs are modelled on the upper layer of the SIW cavity. A feed forward multilayer perceptron (FF-MLP) neural network is used to optimize the physical dimensions of the proposed filters. To validate the analytical results, physical prototypes of the proposed filters are fabricated, and a measurement is carried out with a Combinational Network Analyzer (Anritsu-MS2037C), and the obtained experimental results agree well with the estimated results using full wave analysis. Within the passband from 8.22 to 8.95 GHz, S12 of the first filter shows better than -0.5 dB insertion loss (IL) and a fractional bandwidth of 8.5%, and within the passband from 8.21 to 8.73 GHz, the second filter shows IL about -0.8 dB and a fractional bandwidth of 6.1%.