引言森林火灾是一种突发性强、破坏性大、处置救助较为困难的自然灾害,对森林、森林生态系统和人类带来难以弥补的危害。面对当前环境污染、气候变暖、森林覆盖面积逐年减少的局面,保护森林资源势在必行。森林火灾的相关因素很多,是气候、地形、火源和可燃物诸因子相互作用形成的复杂现象,通常具有复杂的随机性和非线性行为。而神经网络是一种自组织、自学习、容错能力强的高度非线性系统,本文拟采用广义回归神经网络GRNN,利用某森林地区天气因素和FWI(fire INTRODUCTION Forest fires are natural disasters that are sudden, destructive and difficult to handle and rescue. They bring irreparable harm to forests, forest ecosystems and human beings. Faced with the current situation of environmental pollution, climate warming, forest cover area decreased year by year, the protection of forest resources is imperative. Forest fires have many related factors and are complicated phenomena caused by the interaction of climate, topography, fire source and combustibles. They usually have complicated randomness and non-linear behavior. The neural network is a highly nonlinear system with self-organizing, self-learning and fault-tolerant ability. In this paper, the generalized regression neural network GRNN is proposed. By using the weather factors in a forest area and the FWI