The invention discloses a mental workload assessment method based on multi-physiological parameter method integration. The method includes the steps of 1, designing controlled variable experiments oriented to five characters of a target complex task, acquiring functional physiological data of the experiments, and preprocessing the acquired functional physiological data; 2, completing the extraction of sensitive physiological characteristics reflecting the changes of different characters on mental workloads from the physiological data; 3, screening the physiological characteristics to remove interference information in the physiological characteristics; 4, integrating physiological indexes under different characters to obtain a comprehensive physiological parameter characteristic system; 5,according to the physiological characteristics under different characters, conducting regression on the physiological data under the task characters by machine learning; 6, according to the comprehensive physiological indexes, describing the mental workload level of a current operator. By means of the method, the operation result can be directly evaluated, and high real-time performance is achieved; besides, the accuracy is high, and the status of task execution personnel can be monitored in real time.