针对过程控制中工程化建模方法 ,得到的对象模型为传递函数形式 ,而目前的研究是化为状态空间表示 ,因而迭代学习算法及收敛性条件是以矩阵的形式表示 ,不太适合于实际应用的事实。文章用时频结合的方法给出了设定值变化下的迭代学习算法 ,并以系统参数形式地给出几种学习算法的收敛性定理 ,同时通过仿真表明其正确性 For the engineering modeling method in process control, the object model obtained is in the form of transfer function, and the current research is turned into the state space representation. Therefore, the iterative learning algorithm and the convergence condition are expressed in the form of matrix, which is not suitable for the actual The fact of application. In this paper, the iterative learning algorithm with the change of setpoint is given with the combination of time-frequency method. The convergence theorem of several learning algorithms is given in the form of system parameters, and the correctness