Abstract: Some similar seismic anomalies appeared in different time and different areas before Wenchuan M8.0 and Kunlunshan M8.1 earthquakes, which may bring us some clews to the prediction of large earthquakes. ① Two seismic strengthening areas of moderate and large earthquakes formed in Qinghai-Tibet block ten or more years before the two large earthquakes, and the two large earthquakes occurred in the gap in the strengthening area; ② Several years prior to the two earthquakes, huge seismic belts of moderate and large earthquakes formed and the two earthquakes took place in the quiescence area between the huge seismic belts. In the same period, there formed a doughnut-shape pattern of M≥4.0 earthquakes, in which the anomalies of earthquake frequency, load/unload response ration and seismic inhomogeneous degree were outstanding and varied with time, these may be clews to the judgment of middle-term earthquake prediction; ③ The activity of earthquake swarms and clusters is also outstanding. Four earthquake clusters appeared around the Kunlunshan earthquake and the Wenchuan earthquake occurred in the gap formed by earthquake swarms. These may be signals for judgment of middle-term and short-term earthquake prediction; ④ Some small to moderate earthquakes occurred in or near the seismogenic fault zone several months before the two large earthquakes. Ten months prior to Wenchuan earthquake, in the north part of the Longmenshan fault zone, two Qingchuan M≥4.0 shocks and one Songpan M4.3 earthquake occurred,and in the south part, three M≥4.0 earthquakes took place in Kangding and its vicinity, and a small earthquake swarm occurred in Zipingpu reservoir area three months before the Wenchuan earthquake. About one year before Kunlunshan earthquake, a M6.6 earthquake occurred in Xinghai area and a M5.1 shock occurred about 30 km away from the epicenter of Kunlunshan earthquake.
The features above may have some implications. ① The spatial-temporal characteristics of large earthquakes are different from those of moderately strong earthquakes. The present system of earthquake monitoring and prediction does not suit to the anomaly features of great earthquakes; ② Prediction of large earthquakes can not rely singly on seismic precursors and seismic precursors, geological structure and deep exploration should be combined together; ③ A new strategy of earthquake prediction, observation and administration should be developed to suit large earthquake predictions.

Key words: Seismic strengthening area, Seismic doughnut-shape pattern, Quiescence area, Seismic belt, Seismic gap, Earthquake swarm, Earthquake cluster