介绍了一种自动调整遗忘因子的最小二乘方法，并将它引入到ＬＫＬ（最小二乘，Ｋａｌｍａｎ滤波，最小二乘）方法中．大量数字仿真表明这种方法对于小样本快速时变系统的辨识是有效的．自动调整遗忘因子的ＬＫＬ辨识方法应用到灯泡销售预测中，获得较为满意的结果． A least squares method to automatically adjust the forgetting factor is introduced and introduced into LKL (least squares, Kalman filtering, least squares) method. A large number of numerical simulations show that this method is effective for the identification of small sample fast time-varying systems. The LKL method, which automatically adjusts the forgetting factor, is applied to the bulb sales forecast to get satisfactory results.