本篇文章利用2009年至今全国5级以上地震数据，利用K-means聚类算法将全国分成了四个地震区域，把地震当作在这些地震区域间随机游动的质点，且其转移不依赖于两次以前发生的地震。统计各个区域地震发生的频率，计算出地震在各个区域之间的转移概率。利用地震的一阶以及二阶转移概率，建立预报测度用以预测地震发生在各个区域的可能性。利用此预报测度，本文预测了国内2020年7月26日至2021年3月24日发生的11处5级以上地震发生的区域，其中有8次发生在预测的第一可能区域，2次发生在预测的第二可能区域，仅有1次错报，预测结果较好。

This article uses the national earthquake data of magnitude 5 and above from 2009 to the present, and uses the K-means cluster algorithm to divide the country into four earthquake regions. The earthquake is regarded as the mass point，randomly walking between these earthquake regions, and their transfer does not depend on earthquakes occurred twice before. We count the frequency of earthquakes in the divided regions and calculate the transition probability of earthquakes between regions. Using the first-order and second-order transition probabilities of earthquakes, a prediction measure is established to compare the possibility of earthquakes in various regions. Using this prediction measure, this article predicts 11 areas where earthquakes of magnitude 5 or higher occurred in China from July 26, 2020 to March 24, 2021, of which 8 occurred in the most possible area and 2 occurred in the second possible area, there was only one error.