Second-generation antipsychotics (SGAs) are widely used to treat schizophrenia (SCZ), but they often induce metabolic side effects like dyslipidemia and obesity. We conducted genome-wide association studies (GWASs) to identify genetic variants associated with SGA-induced lipid and BMI changes in Chinese SCZ patients. A longitudinal cohort of Chinese SCZ receiving SGAs was followed for up to 18.7 years (mean = 5.7 years, SD = 3.3 years). We analysed the patients' genotypes (N = 669), lipid profiles, and BMI using 19 316 prescription records and 3 917 to 7 596 metabolic measurements per outcome. Linear mixed models were employed to evaluate seven SGAs' random effects on metabolic changes for each patient, followed by GWAS and gene set analyses with Bonferroni and FDR correction. Five SNPs achieved p-value < 5 x 10(-08) before multiple testing correction: rs6532055 (ABCG2) linked to olanzapine-induced LDL changes, rs2644520 (near SORCS1) linked to aripiprazole-induced triglyceride changes, rs115843863 (near UPP2) linked to clozapine-induced HDL changes, rs2514895 (near KIRREL3) linked to paliperidone-induced LDL changes, and rs188405603 (SLC2A9) linked to quetiapine-induced triglyceride changes. These five SNPs passed FDR correction at 0.2 but not Bonferroni-corrected genome-wide significance threshold (p-value < 3.125 x 10(-10)) for 160 GWAS analyses. Gene-based analysis revealed six genome-wide significant genes after Bonferroni correction (p-value < 2.73 x 10(-6)): ABCG2, APOA5, ZPR1, GCNT4, MAST2, and CRTAC1. Four gene sets were significantly associated with SGA-induced metabolic side effects. In summary, this pharmacogenetic GWAS identified several genetic variants potentially associated with SGA-induced metabolic side effects, potentially informing personalized treatment strategies to minimize metabolic risk in SCZ patients. Given our limited sample size, further replications are required to confirm the findings.