Abstract: The Acceleration Index (AI) algorithm was applied in a statistical retrospective research for earthquakes with M_L≥6.0 since 1970 in northern China.We calculated the activity frequency of small-middle earthquakes before strong earthquakes with this method. The results show that AI vales can reflect abnormal frequency of mid-small earthquakes in space. Moreover, studying the seismicity of small-middle earthquakes before and after the moderately strong earthquakes (M_L≥6.0) with Frequency-Area Method (namely S-value method), this method gives support on the calculating results of AI algorithm in time. AI valuse can quantitatively indicate activity strengthening or weakening level of micro-earthquakes before the moderate-to-strong shocks. The S-value of moderate-strong earthquakes exhibits high anomalous values or low anomalous ones before strong earthquake occurs and it decreases or increases after the strong earthquake occurred. Abnormal point of S-values and abnormal result of AI scan values are complementary in space and time. In the case of integral earthquake catalog, small-middle earthquakes activity 1 year before the earthquake compared with the activity background of previous three years, five years in moderately strong earthquake epicenter and its adjacent area, it shows obviously “accelerated” or “slow down” phenomenon. The change of the two parameters can be used as indicators for moderate-strong earthquake prediction in north China.

Key words: AI algorithm, Seismic activity, Earthquake prediction, North China