In 2017, a new two-step algorithm for the treatment of COPD was proposed. This algorithm was based on the severity of symptoms and phenotypes or treatable traits, and patient-specialised assessment targeting eosinophilic inflammation, chronic bronchitis, and frequent infections is recommended after exacerbation occurs despite maximal bronchodilation therapy. However, recent studies have revealed the clinical characteristics of patients who should have second controllers added, such as ICS. We again realized that treatable traits should be assessed and intervened for as early as possible. Moreover, the treatment algorithm is necessary to be adapted to the situation of clinical practice, taking into account the characteristics of the patients. The time to revise COPD treatment algorithm has come and we propose a new 3-step parallel approach for initial COPD treatment. After the diagnosis of COPD, the first assessment is to divide into two categories based on the usual clinical characteristics for patients with COPD and the specific clinical characteristics for each patient with concomitant disease. In the former, the assessment should be based on the level of dyspnea and the frequency of exacerbations. After the assessment, mono- or dual bronchodilator should be selected. In the latter, the assessment should be based on asthma characteristics, chronic bronchitis, and chronic heart failure. After the assessment, patients with asthmatic characteristics may consider treatment with ICS, while patients with chronic bronchitis may consider treatment with roflumilast and/or macrolide, while patients with chronic heart failure may consider treatment with selective β1-blocker. The 3-step parallel approach is completed by adding an additional therapy for patients with concomitant disease to essential therapy for patients with COPD. In addition, it is important to review the response around 4 weeks after the initial therapy. This COPD management proposal might be considered as an approach based on patients' clinical characteristics and on personalized therapy.