The invention discloses a metabonomics data fusion optimization processing method. The method comprises the following steps of 1, converting metabonomics data of many patients of different sources into metabonomics data matrixes separately, 2, using an experiment design method to optimize a metabonomics data optimal preprocessing method with different sources separately, then through the optimized data optimal preprocessing method to process the metabonomics data matrixes in the step 1, and through the combination of a classification model, finding metabolite data with great contribution; 3, converting the metabolite data with the great contribution found in the step 2 into an important metabolic data matrix; 4, using the experiment design method to optimize the data optimal preprocessing method of the important metabolic data matrix, through the optimized data optimal preprocessing method, processing the important metabolic data matrix obtained in the step 3, and through the combination of the classification model, typing and classifying different patients. According to the metabonomics data fusion optimization processing method, the accuracy for typing and classifying diseases is improved, and a doctor can be assisted in more precisely laying out a personalized treatment scheme for a patient.