The invention relates to an online monitoring data processing method.The method comprises the steps of 1, for monitoring data corresponding to a certain online monitoring parameter acquired continuously, reading the sequence of the stored online monitoring data x(n)=x1, x2, ..., xn at intervals of T, comparing the size of n data to obtain the maximum value xmax and the minimum value xmin, and dividing the section [xmin, xmax] into at least two sections; 2, conducting statistics on the number of data from the sequence x(n) in each section, and then calculating the probability of data in each section; 3, taking the section with the highest probability as a confidence interval, and calculating the average value of data in the confidence interval, wherein the average value is the monitoring effective value used for storage and forward.By the adoption of the method, unreasonable data uploaded by a monitoring device can be filtered out, useful feature information in the monitoring data can be preserved to the maximum extent, and the processed online monitoring data has higher stability and practical value.