In order to predict the changing trend of grinding force in ultrasonic grinding of alumina ceramics,a method based on the combination of fuzzy information granulation and support vector machine was proposed.Firstly,the ultrasonic grinding test of alumina ceramics was carried out.The grinding force obtained by the experiment was granulated by fuzzy information granulation method.The artificial immune system algorithm and the particle swarm algorithm were mixed in parallel to form the artificial immune system particle swarm algorithm (AISPSO).Secondly,a nonlinear regression support vector machine prediction model was established and fuzzy particles were predicted.The support vector machine prediction model was optimized through the AISPSO algorithm.Finally,the change trend and range of grinding force in ultrasonic grinding of alumina ceramics were obtained.The results show that this method can effectively realize the prediction of the change trend and range in ultrasonic grinding.And the error of prediction of the change range of the next 5 sets of data are less than 10%.A new idea for adjusting the process parameters by changing the grinding force to obtain a better-processed surface is provided.