Nowadays, in forensic laboratories, heroin profiling is frequently carried out by gas chromatography coupled with mass spectrometry (GC-MS). This analytical technique is well established, provides good sensitivity and reproducibility, and allows the use of large databases. Despite those benefits, recently introduced analytical techniques, such as ultra-high-pressure liquid chromatography (UHPLC), could offer better chromatographic performance, which needs to be considered to increase the analysis throughput for heroin profiling. With the latter, chromatographic conditions were optimized through commercial modeling software and two atmospheric pressure ionization sources were evaluated. Data obtained from UHPLC–MS/MS were thus transferred, thanks to mathematical models to mimic GC-MS data. A calibration and a validation set of representative heroin samples were selected among the database to establish a transfer methodology and assess the models' abilities to transfer using principal component analysis and hierarchical classification analysis. These abilities were evaluated by computing the frequency of successful classification of UHPLC–MS/MS data among GC-MS database. Seven mathematical models were tested to adjust UHPLC–MS/MS data to GC-MS data. A simplified mathematical model was finally selected and offered a frequency of successful transfer equal to 95%.