Abstract: Utilizing more than 80 earthquake sequences with magnitude Ms ≥ 5.0 which are classified into three types:main shock-after shock type, strong earthquake swarm type (including double main earthquake type) and isolated shock type by traditional classification method given by reference1, the relationship between the first big earthquake M₁ and the largest succeding eathquake M₂ has been studied, and the linear equation M₂=a+bM₁ and a, b value have been obtained for different types respectively.The comparison between M₂ calcuated from the equation derived in this paper and M₂ in the real sequence shows that the accuracy Of the internal coincidence test exceeds 90% under the standard|△M| ≤ 0.5. It is proved that the accuracy is adout 90% for the extrapolation test using medium-small earthquake sequences and historical or modern medium-strong earthquake data. So this method shows relatively high practiCal value in post-shock trend prediction and the estimation of the largest succeeding earthquake in field study, To appleing this method conveniently to early trend diagnoses in field study, the synthesized classification indexes of the earthquake sequences have been achieved from the sequences data within 24 hours after M₁ had drawn form research pattern recognition method.