We use a Probabilistic Neural Network (PNN) technique to derive the orbital parameters of spectroscopic binary stars. Using measured radial velocity data of five double-lined spectroscopic binary systems (i.e., EQ Tau, V376 And, V776 Cas, V2377 Oph and EE Cet), we find the corresponding orbital and spectroscopic elements. Our numerical results are in good agreement with those obtained by other groups via more traditional methods.