Introduction: Wheelchair Users (WCUs) depend on their upper extremities for their daily living. Therefore, it is not unusual to find that shoulder pain (SP) is a problem for WCUs and reduces their participation in sport and leisure activities. Objectives: The aims of this study were 1 – to analyse skin temperature measured by infrared thermography (IRT) before (pre-test), one minute after (post-test) and 10 min after (post-10) the kinematic wheelchair propulsion test (T-CIDIF) of athletic wheelchair users; 2 – to evaluate the relationship between shoulder pain (SP) and Skin Temperature Asymmetry (ΔTsk) before and after (pre-test, post-test, post-10) the T-CIDIF, and to relate the SP with the kinematic variables of the T-CIDIF. Participants & interventions/procedure: A volunteer sample of 12 wheelchair athletes completed an exercise test (T-CIDIF) in their own wheelchair. It consisted in a 30-s maximum test performed on two rollers. Two linear transducers connected to the rollers registered the number of propulsions, maximum and mean velocity and power of each arm. SP was assessed with the Wheelchair Users Shoulder Pain Index (WUSPI). Skin temperature (Tsk) of the anterior and posterior upper body was measured before and after the T-CIDIF by using an infrared camera. A total of 26 ROIs were evaluated with respect to the opposite side of the body to identify significant (ΔTsk). Results/main outcome measure(s): Significant differences were observed between the Tsk of the post-10 and pre-test in 12 ROIs, and between the post-10 and the post-test in most of the ROIs. These differences are attenuated when the ΔTsk is compared before and after exercise. Tsk tends to initially decrease immediately after the test and then significantly increase after 10 min of completing the T-CIDIF. The ΔTsk vs SP analysis yielded significant inverse relationships (from r = −0.58 to r = −0.71, p < 0.05) in 5 of the 26 ROI. No significant correlations between propulsion variables and SP questionnaire were found. All T-CIDIF variables were significantly correlated with the temperature asymmetries in multiple ROIs (from r = −0.86 to r = −0.58, from p < 0.05 to p < 0.001). Conclusions: These results present indications that high performance wheelchair athletes exhibit similar capacity of heat production than able-bodied. The thermographic data inversely correlates with the SP and the kinematic variables, but the last is not related to SP. This work contributes to improve the understanding about temperature changes in wheelchair athletes during exercise, and could be used to assess the efficacy of various sports and rehabilitation programs.