BACKGROUND Stroke is a common cause of neurological dysfunction, often resulting in hemiplegia. Thus, rehabilitation of limb function in stroke patients is an important step towards accelerating recovery and improving quality of life. OBJECTIVE To investigate whether unilateral strength training in hemiplegic stroke patients could lead to cross-migration, an increase in bilateral muscle strength, and an improvement in lower limb motor function. METHODS We randomly divided 120 patients with hemiplegic stroke into two groups: Eexperimental and control groups, with 60 patients in each group. Both groups received routine rehabilitation treatment, while the experimental group additionally received ankle dorsiflexion resistance training on the healthy side for 6 wk. We measured the maximum voluntary contract (MVC), changes in surface electromyography (EMG), and the lower limb motor function using the simplified Fugl Meyer Motor Function Assessment Scale (FMA) before and within 1 wk after training. RESULTS The FMA score in the experimental group improved significantly compared to both their pre-training score and the control group's post-training score (P < 0.05). The integrated EMG of the anterior tibialis muscle and pulmonary intestine muscle in the experimental group were significantly different after training than before (P < 0.05). Furthermore, the MVC of the anterior tibialis muscle on both the healthy and affected sides and the MVC of the pulmonary intestine muscle on both sides showed significant improvement compared with before training and the control group (P < 0.05). CONCLUSIONS Our findings suggest that ankle dorsiflexion resistance training on the healthy side in hemiplegic stroke patients can increase strength in the opposite tibialis anterior muscle and antagonist's muscle, indicating a cross-migration phenomenon of strength training. Furthermore, this type of training can also improve lower limb motor function, providing a new exercise method for improving early ankle dorsiflexion dysfunction.