Introduction. This study investigates the variations of nonlinear parameters of heart rate variability (HRV) due to an arithmetic stress test. Material and methods. We tested 83 healthy students with a mean age of 21.12±0.21 years. The students were examining in the following sequence: recording of HRV at rest (10 minutes), during mental arithmetic (10 minutes) and during the recovery period (10 minutes). HRV was assessed by applying Poincare plot analysis (SD1, SD2, SD2/SD1, SS) to the R-R interval. The traditional (GI) and redefined Guzik’s index (GIp) (Karmakar, 2009) was calculated for each subjects in each phase of the experiments. Results and discussion. SD1 during mental arithmetic was significantly smaller (24.81±1.80) than that at rest (34.90±2.08; p<0.0001). During recovery from mental arithmetic tasks SD1 return to baseline level (39.01±3.65). SD2 tended to decrease during the mental arithmetic task, reaching its highest levels during the period of recovery (from 70,74±2,94 during tasks to 81,51±4,23 during post stress; p<0,004). SD1/SD2 ratio during mental stress (0.333±0.013) was significantly lower than that at rest (0.474±0.017; p<0.00001) and during the period of recovery (0.441±0.017; p<0.00001). SS was significantly lower during the test (2802.32±320.04) than that at rest (3860.98±377.56; p<0.001) and recovery phase (5812.81±971.32; p<0.001). There was no significant difference in the GI during mental stress (0.491±0.004) than that at rest (0.496±0.004; p>0.05) and recovery period (0.486±0.004; p>0.05). After redefinition traditional asymmetry indices, we have found pronounced changes in GIp parameter with a maximum at rest and minimum during arithmetic test (0.436±0.012 vs 0.478±0.013; p<0.05). This findings suggest that nonlinear HRV analysis could be effective in automatically detecting functional status during mental stress.