Torrefaction is a crucial biomass upgrading technology to produce biochar for fuel, soil amendment, and bioabsorbent. A number of indicators such as weight loss (WL), torrefaction severity index (TSI), and severity factor (SF) have been conducted to describe the torrefaction degree. However, operating conditions such as torrefaction temperature and duration are not considered in weight loss and torrefaction severity index, while biomass nature is not taken into account in the severity factor. To overcome these drawbacks, an indicator termed torrefaction severity factor (TSF) is proposed by introducing a time exponent in the severity factor. Four different biomass materials of Chinese medicine residue, Arthrospira platensis residue, C. sp. JSC4, and spent coffee grounds are examined. After the optimization of the time exponent, TSF can accurately correlate weight loss and thereby torrefaction severity, and improve the prediction up to 13% when compared to severity factor. In addition, the results suggest that TSF is able to appropriately predict the enhancement factor of HHV and energy yield where the coefficient of determination (R-2) is beyond 0.83. Overall, TSF has successfully combined the operating conditions (temperature and duration) and biomass species, can be utilized for predicting torrefaction performance. This gives a simple and fast way for torrefaction operation and reactor design, thereby achieving time-saving and efficient predictions.