A method of detecting process abnormal signals using a recursive maximum likelihood method and Bayesian statistical inference has difficulty mselecting the order of an AR model. A method using process signals without trend ingredient by once differential method and Bayesian statistical inference has been developed. The method proposed in this study has the advantage of detecting online process abnormal signals in industrial use. It was applied to abnormal detection of the catalyst feed flow in a linear low-density polyethylene plant to confirm the design philosophy. The actual result indicates that the proposed method is effective in detecting abnormal process signals.