In order to evaluate the health status of the equipment on-line, an improved Mahalanobis distance combined with sliding window evaluation method is proposed. The method by monitoring the time series of multiple related performance parameters, on the basis of Mahalanobis distance, using improved Mahalanobis distance method will obtain the mean deviation and outlier detection from the rated parameters of equipment instead of collecting samples. In order to remove transient faults and environmental disturbance, increase the robustness of the detection, the improved Mahalanobis distance time series based on histogram method and the sliding window contains the parameters of historical information, construct degradation index, evaluation of equipment state of health. A case study of an air compressor in an oxygen plant is given to demonstrate the feasibility and effectiveness of the method.