Cognitive diagnostic assessment seeks to promote targeted instruction designed to address learners' strengths and weaknesses within a specific domain. Most of the previous CDA applications, however, have merely implemented a single CDM. Few studies have investigated the applicability of multi-CDM to educational assessment data. To this end, the present study attempted to diagnose a sample of 740 college freshmen's performance on a reading comprehension test designed by the PELDiaG (Personalized English Learning: Diagnosis & Guidance) research team from a key university in China with the multi-CDM. The model-data fit results showed that the multi-CDM outperformed any single CDMs and enhanced the interpretation of the inter-skill relationship as well. Consequently, both group and individual level diagnostic information were extracted and further synthesized into a fine-grained diagnostic feedback. The findings provided further evidence pertinent to the practicality of multi-CDM in reading comprehension tests. Finally, limitations and suggestions for further research were presented.