This study identified several physiological indices that can accurately monitor mental workload while participants performed multiple tasks with the strategy of maintaining stable performance and maximizing accuracy. Thirty male participants completed three 10-min. simulated multitasks: MATB (Multi-Attribute Task Battery) with three workload levels. Twenty-five commonly used mental workload measures were collected, including heart rate, 12 HRV (heart rate variability), 10 EEG (electroencephalography) indices (alpha, beta, theta, alpha/theta, theta/beta from O1-O2 and F4-C4), and two subjective measures. Analyses of index sensitivity showed that two EEG indices, theta and alpha/theta (F4-C4), one time-domain HRV-SDNN (standard deviation of inter-beat intervals), and four frequency-domain HRV: VLF (very low frequency), LF (low frequency), %HF (percentage of high frequency), and LF/HF were sensitive to differentiate high workload. EEG alpha/theta (F4-C4) and LF/HF were most effective for monitoring high mental workload. LF/HF showed the highest correlations with other physiological indices. EEG alpha/theta (F4-C4) showed strong correlations with subjective measures across different mental workload levels. Operation strategy would affect the sensitivity of EEG alpha (F4-C4) and HF.