本文提出一种由多个既关联又独立的子网络构成的用于逼近非线性函数的多值学习网络（KLN），并在此基础上，研究了具有自学习功能的多值逻辑控制．网络运算主要为整数的加法和逻辑判断．仿真结果表明，基于KLN的多值逻辑控制系统对复杂非线性系统具有良好的学习控制性能，且具有计算简单和省时的特点． In this paper, we propose a multi-valued learning network (KLN) composed of multiple correlated and independent sub-networks for approximating non-linear functions. On this basis, we study a multi-valued logic control with self-learning function. Network operations are mainly integer addition and logic. The simulation results show that the KLN-based multi-valued logic control system has a good learning control performance for complex nonlinear systems, and has the characteristics of simple calculation and time-saving.