Excessive mental workload represent a critical risk factor for workplace accidents. Heart rate variability (HRV) is a non-invasive low cost electrophysiological autonomic biomarker related to emotional and cognitive regulation. Several studies report that mental overload impairs parasympathetic-mediated HRV indices (e.g. rMSSD). However, the influence of resting state HRV as a predictor of long-term mental workload impairments remains unknown. Thirty participants (22 males; 8 females) had their HRV measured (5-min period) before performing the number search task. After the task, the mental load was accessed by the NASA-TLX questionnaire. A simple linear regression model between HRV and NASA-TLX dimensions showed that resting state rMSSD is associated to physical demand (ND-2, R = 0.143, p = 0.03) and frustration level (ND-6, R = 0.175, p = 0.02) dimensions of mental workload. The comparison between 1 and 5-min epochs suggests that regression models remain reliable even using the ultra-short term HRV (< 1 min) recording values (R values from 0.11 to 0.15 for ND-2 and R values from 0.16 to 0.19 for ND-6). These results suggest that resting state HRV is associated to long-term effects of mental workload on physical and emotional demands. In addition, the ultra-short term HRV indices remains reliable to assess ND-2 and ND-6 dimensions of mental workload when compared to gold-standard time interval (> 5 min). The resting state cardiac autonomic tone assessment optimizes the physiological approach with a quick, non-invasive and low-cost assessment that can provide insights about mental load adjustments to prevent work-related accidents.