To obtain rich information about the cognitive diagnosis of borderline personality disorder (BPD), this study attempted to retrofit a traditional borderline personality questionnaire so that the improved assessment (called CDA-BPD) could provide more diagnostic information. The retrofitting processes included the following steps: (1) applied an cognitive diagnosis model to analyze the psychometric characteristics of the traditional questionnaire; (2) under the guidance of cognitive diagnosis assessment (CDA), high-quality items were chosen to develop the CDA-BPD and tested on 1,097 subjects; (3) the quality of the CDA-BPD was evaluated; (4) the structure of the CDA-BPD was analyzed. Results indicated that: (1) the CDA-BPD had acceptable reliability and validity; (2) the CDA-BPD had sensitivity of 0.985 and specificity of 0.853 with area under curve (AUC) = 0.956; (3) the two structural factors of the traditional questionnaire were confirmed in the CDA-BPD; χ2 was 83.01 with df = 26, p < .0001, comparative fit index (CFI) = 0.97, root mean square error of approximation (RMSEA) = 0.045. It was concluded that the practice of retrofitting a traditional borderline personality assessment for cognitive diagnostic purpose was feasible. Most importantly, under the cognitive diagnosis model framework, CDA-BPD could simultaneously provide general-level information and the detailed symptom criteria-level information about the posterior probability of satisfying each symptom criterion in the Diagnostic and Statistical Manual of Mental Disorders (5th edition; DSM-5; American Psychiatric Association, 2013) for each individual, which gave further insight into tailoring individual-specific treatments for borderline personality disorder.