Purpose: This study aimed to examine the effects of proprioceptive neuromuscular facilitation (PNF) lower extremity patterns combined with elastic bands applied to stroke patients diagnosed with hemiplegia through self-training using the non-paralyzed side approach and the paralyzed side approach, and to investigate the differences in the effects. Methods: Nine chronic stroke patients who were being treated not more than twice a week at H, K, R, and C hospitals located in Gangwon-do, performed self-training for 16 minutes, two times per day for four weeks between August and October 2015. The subjects' balance ability was measured using the Tetrax stability index, the weight distribution index, the Berg Balance Scale (BBS), and the Timed Up and Go (TUG) test before the experiment and four weeks later. Among the statistical methods, paired t-tests were conducted for intra-group comparison of the measurements taken before and after the experiment, and independent t-tests were conducted for inter-group comparison of the ex post facto values. The statistical significance level was set to 0.05. Results: When the lower extremity patterns were applied to the non-paralyzed side group and the paralyzed side group, significant intra-group differences were observed for the Tetrax stability index, the weight distribution index, the Berg Balance Scale (BBS), and Timed Up and Go (TUG) tests (p<0.05); however, the inter-group comparisons showed no significant differences. Conclusion: The non-paralyzed side approach was found to be easy for patients to participate in and it also affected the patients' paralyzed side. Although the paralyzed side approach produced good exercise effects in a short period of time, it could result in adverse effects, such as a decrease in motivation and self-confidence. Therefore, these approaches are considered to be more effective when they are selectively applied depending on the purpose of the intervention and the degree of a patients' participation.