Individuals with stroke often have impaired paretic arm function. While alterations in the paretic limb’s movement during tasks like reaching and grasping have been noted, these do not fully capture the motor behavior in transport tasks. This study aimed to examine the movement of paretic and non-paretic limbs during a mug transportation task using kinematic metrics and assess their correlation with a transport clinical test. Twenty-two individuals with chronic hemiparesis from a stroke were asked to transport a mug across a table using their non-paretic limb and then with their paretic limb. The kinematic analysis compared accuracy, precision, movement time, peak velocity, motor control strategy, efficiency, and interjoint coordination between limbs. The box and block test was used for correlation analysis. Results indicated the non-paretic limb had higher precision (∼52 %;​ forward end-range:​ 10.7 ± 5.7 mm) and efficiency (∼34 %;​ object path ratio:​ 1.38 ± 0.14 index) than the paretic limb (16.3 ± 13.5 mm and 1.51 ± 0.26 index, respectively), with no significant differences in other kinematic metrics. A strong correlation was found in the paretic limb between the box and block test and efficiency (r = −0.75), and good correlations with movement time (r = −0.66) and smoothness (r = −0.69). Despite the relevance of other kinematic metrics, efficiency was key for the transport task and a major factor in motor impairment. Rehabilitation for post-stroke individuals should thus emphasize transport tasks, aiming to increase efficiency.