Disclosed are a method and device for detecting an abnormal load. The method includes: collecting data of each load item in all hosts; pre-processing to obtain a second sequence corresponding to each load item; obtaining a coefficient matrix and a detail vector; computing a weighted mean of abnormal load probabilities of all the coefficient vectors, to obtain an abnormal load probability of each second sequence; comparing the abnormal load probability with a confidence interval, and judging whether there is an abnormal load, and judging that there is no abnormity in the corresponding second sequence if the abnormal load probability of each second sequence is in the confidence interval; judging whether there is an abnormal load according to data of all load items and the abnormal load probability of each second sequence; and finding a bearing host that has the abnormal load from the current online service.