This study aimed to assess the effect of bull exposure to high temperature-humidity index (THI) on spermatozoa kinetics variables. In partnership with a commercial bull semen processing stud, we used data of 2,758 semen batches from 57 bulls. Semen samples were collected by artificial vagina between July/2017 and July/2020. Bulls were housed at Uruguaiana, RS, Brazil (29° 45’ 42.884” S 57° 5’ 9.034” W), in individual stalls, fed twice daily according to the nutritional demand for bulls in a twice a week semen collection program. Animals had access to ad libitum water and shade, but no additional temperature or humidity control. Weather data were obtained from a public weather station located approximately 30 km away from the bull stud. We tested 25 statistical models for each dependent variable considering five THI equations and five THI variables. Bull was included as covariate in all models. The statistical model using the number of days when THI reached 74 or above, calculated by the Thom equation (Cooling degree-days, Weather Bureau, 1958) best explained (higher determination coefficient) the changes in all sperm kinetics (data not shown). Then semen batches were separated by quartiles according to the number of days when the THI was greater than 74 (THI>74) during the 60-day window prior to the semen collection day. Quartiles (Q) 1, 2, 3 and 4 represent 0 to 9, 10 to 21, 22 to 41, and more than 42 days of exposure to THI reaching 74 or higher values, respectively. The effect of the quartiles on the dependent variables pre-freezing motility (%), mobile cells (%), motility loss (percentage points), progressive cells (%), fast cells (%), concentration, VSL (mm/s), VCL (mm/s), VAP (mm/s), LIN (%), STR (%), and ALH (mm) was assessed by mixed models using bull as random effect. A significant Q effect, where all Q were different from each other, was observed on motility loss (greater Q = greater increase), mobile cells and fast cells (greater Q = greater reduction). There were also Q group differences for progressive cells and semen concentration, with Q4 reaching lower values than Q1 and Q2. Quartile 4 had lower values of speed variables than Q1 and Q2 (VAP, VCL, VSL), whereas rectilinearity variables (LIN, STR) demonstrated that Q4 samples had greater % of cells with a more rectilinear path. Quartile 4 cells had lower amplitude of movement from their projected real path (ALH) than Q1 and Q2. Quartiles had no effect on pre-freezing motility. In conclusion, a greater the number of days of bull exposure to THI>74 during the 60 days prior to semen collection seems to lead to a greater negative impact on post-freezing sperm cells kinetics. Overall, the data suggest functional impairment of the sperm cells as bulls are exposed to a THI associated with heat stress.