This paper proposes an inhaler-use behavior evaluation method that uses inertial sensors to monitor patients with bronchial asthma and chronic obstructive pulmonary disease (COPD). COPD is common disease, and the accurate use of inhalers is vital to controlling its symptoms. However, many patients improperly use their inhalers. Hence, by augmenting an Ellipta™ inhaler with inertial measurement units, this study evaluates patient inhalation motions using the acquired motion data. Compared with conventional methods, ours is less affected by external factors, such as sound and temperature, and it can be applied outside clinical settings. Specifically, a linear discriminant analysis algorithm is provided for selecting characteristic variables for each error type, and a judgment method using these variables is proposed. A dynamic programming matching algorithm is then applied to determine correctness. Experimental results show that our method provides high discriminant accuracy, and it accounts for the similarity of waveforms, which allows us to visualize errors, unlike contemporary methods. We expect that our inhalation method and the accompanying dataset will offer valuable guidance for future research as well as useful feedback to patients.