Abstract: Noise can be used in array bistable stochastic resonance (SR) systems to furtherly enhance the ability of weak signal detection over isolated one. This leads to new methods of weak signal detection from heavy background noise. In this paper, we investigate a method of weak signals detection Noise can be used in array bistable stochastic resonance (SR) systems to furtherly enhance the ability of weak signal detection over isolated bistable one. This leads to a new method of weak signal detection from heavy background noise. In this paper, we investigate a method of weak signal detection using array bistable SR theory. By introducing stationary auto-covariance, we study the signal-to-noise ratio (SNR) gain by using both theoretical and numerical arguments. Based on this study, we analyze the effects of array noise, external noise and array sizes on detection performance. Finally we compare the corresponding detection method with the one using isolated bistable SR. Both theoretical analysis and numerical simulation show that the detection performance is much improved by using an array bistable SR than an isolated bistable one under the same conditions. Thus this study is significant for the development of array bistable SR theory and applications.