Anaplastic lymphoma kinase (ALK) is implicated in the genesis of multiple malignant tumors. Lorlatinib stands out as the most advanced and effective inhibitor currently used in the clinic for the treatment of ALK-positive non-small cell lung cancer. However, resistance to lorlatinib has inevitably manifested over time, with double/triple mutations of G1202, L1196, L1198, C1156 and I1171 frequently observed in clinical practice, and tumors regrow within a short time after treatment with lorlatinib. Therefore, elucidating the mechanism of resistance to lorlatinib is paramount in paving the way for innovative therapeutic strategies and the development of next-generation drugs. In this study, we leveraged multiple computational methodologies to delve into the resistance mechanisms of three specific double mutations of ALKG1202R/L1196M, ALKG1202R/L1198F and ALKI1171N/L1198F to lorlatinib. We analyzed these mechanisms through qualitative (PCA, DCCM) and quantitative (MM/GBSA, US) kinetic analyses. The qualitative analysis shows that these mutations exert minimal perturbations on the conformational dynamics of the structural domains of ALK. The energetic and structural assessments show that the van der Waals interactions, formed by the conserved residue Leu1256 within the ATP-binding site and the residues Glu1197 and Met1199 in the hinge domain with lorlatinib, play integral roles in the occurrence of drug resistance. Furthermore, the US simulation results elucidate that the pathways through which lorlatinib dissociates vary across mutant systems, and the distinct environments during the dissociation process culminate in diverse resistance mechanisms. Collectively, these insights provide important clues for the design of novel inhibitors to combat resistance.