An integrated approach combining data acquisition using MS(E) and multi-period product ion scan (mpMS/MS), with high-resolution characteristic extracted ion chromatograms (hcXIC) as a data mining method, was developed for in vivo drug metabolites screening and identification. This approach is illustrated by analyzing metabolites of a potential anticancer agent, 3,6,7-trimethoxyphenanthroindolizidine (CAT) in rat urine based on rapid resolution liquid chromatography combined with tandem mass spectrometry (RRLC-MS/MS). Untargeted full-scan MS(E) enabled the high-throughput acquisition of potential metabolites, and targeted mpMS/MS contributed to the sensitivity and specificity of the acquisition of molecules of interest. The data processing method hcXIC, based on the structure of CAT, was shown to be highly effective for the metabolite discovery. Through the double-filtering effect of the characteristic ion and accurate mass, conventional extracted ion chromatograms that contained a substantial number of false-positive peaks were simplified into chromatograms essentially free of endogenous interferences. As a result, 21 metabolites were detected in rat urine after oral administration of CAT. Based on the characteristic fragmentation patterns of the phenanthroindolizidine alkaloid, the structures of 9 metabolites were identified. Furthermore, the interpretation of the MS/MS spectra of these metabolites enabled the determination of demethylation position as well as the differentiation between N-oxidized and hydroxylated metabolites.