介绍了近年来地震预测学中非线性理论和非线性方法的研究趋势和最新研究进展。在非线性数学模型的研究中，较为成功地将非线性门限模型、指数模型、具有长程关联的自回归模型等应用于地震的中期预测上。在分形研究上，明确地取代了以往的一些经验结果，使对地震的活动的经验统计关系上升到分形的理论。在研究地震的动力学机制中，以多滑块—弹簧模型为代表，模拟地壳上地震断层运动机制和事实，建立了粘滑摩擦情况下的参数动力学方程，方程导出的动力行为出现了极限环、分岔、混沌等非线性行为，使对地震的有物理基础的概率预报有了明确的物理基础。基于地震活动的自组织临界现象，依赖于细胞自动机，较好地模拟出了地震的自组织行为，得到了接近于实际的地震活动统计规律。基于临界现象的重整化群方法应用于地震中期预报有深刻的物理意义。

Advancements of nonlinear science theory and method in seismic forecasting application are indicated and commented. In the respect of nonlinear mathematical models, the threshold regression model, the exponential model, and long-distance-relation ARIMA model .etc., successfully have been applied into the earthquake prediction in the medium term. In the fractals study, some new finding or discovery has been obtained for the previous experienced law or relation of the seismicity. In the dynamic mechanism studying, the multi-spring-block model as a better instance is analogous to the generating mechanism of earthquake by fault motion. For the multi-spring-block model, the dynamic equations with stick-slip friction law have been established and through the equations the typicaly. Nonlinear behaviors, such as limit cycle,bifurcation and chaos etc.,have been derived.The model is useful to development of probability prediction of an earthquake based on physics principle. The self-organization based on seismicity explained well some seismic activity law by means of cellular automata. And the Renormalization Group method for the criticality phenomenon of earthquakes has been applied into the medium-term earthquake prediction with an obviously physical significance.