A relay network can provide the extended link capacity and coverage thanks to its good spatial diversity. In order to analyze the relay network performance, most of the works in literature assume the independent relaying channels. In practice, the received signals experience the common scatterers, hence the correlated relaying channels should be considered. In this paper, we first derive a novel and accurate probability density function (PDF) of a modulus of the sum of products of two correlated zero-mean complex Gaussian random variables (RVs), which is useful for the statistical analysis of the cascaded relaying fading channels. Based on this result, we secondly investigate the received signal-to-noise ratio (SNR) outage probability of a relay network in the presence of correlated relaying channels. It is found that the correlation magnitude is more important than the correlation phase for the analysis of power-based signal-detection of relay networks. Numerical simulations have been carried out to confirm the correctness of the derivation. Index terms-Probability density function, correlated complex Gaussian RVs, relay network.