Introduction The aim of this retrospective cross-sectional study is to examine the routine use proportion and factors determining the use of the Canadian Occupational Performance Measure in the real-world subacute rehabilitation setting. Methods This study retrospectively collected data from all inpatients and occupational therapists at a single Japanese subacute rehabilitation ward during 2017, including Functional Independence Measure motor/cognitive scores, years of experience, and rate of Canadian Occupational Performance Measure administration (that is, Canadian Occupational Performance Measure proportion). Multiple logistic regression analyses were used to identify the determining factors of daily routine Canadian Occupational Performance Measure use/non-use, after which cut-off values were calculated. Results Of the 619 included clients, the Canadian Occupational Performance Measure was applied in 232 cases (37%). A multiple logistic regression analysis revealed two significant determining factors of its use: Canadian Occupational Performance Measure proportion (odds ratio, 1.06) and Functional Independence Measure cognitive item (odds ratio, 1.22). The cut-off value, sensitivity, and specificity, respectively, were 35.4%, 0.73, and 0.36 ( p < 0.0001) for the Canadian Occupational Performance Measure proportion and 25.5, 0.84, and 0.45 ( p < 0.0001) for the Functional Independence Measure cognitive score. Conclusion The client’s high-level cognitive skill and occupational therapist’s attitude may determine the use of the Canadian Occupational Performance Measure.