A method for processing multiple continuous Top-K queries, which is performed between a master server and multiple of slave servers, including steps of: a first step, for arbitrarily selecting the multiple of slave servers and querying and counting up K of accumulated values of which are recorded at most; a second step, for calculating every two adjacent values which have been sent from the same servers to obtain an average value as a threshold; and a third step, for measuring variations of an upper bound and a lower bound for each of the values by using the threshold, and reporting to the master server at a time of the value being in excess of the upper bound or lower than the lower bound for each of the values. |