The spatial arrangement of cardiac myocytes (CMs) and other non-myocytes scaffolds, closely resembling native tissue, is essential to control the CM morphology and function for cardiac tissue regeneration. In the current study, micropatterned fibrous scaffolds were developed to establish a CM co-culture system with cardiac fibroblasts (CFs) and endothelial cells (ECs) as a potential in vitro drug screening model. To pursue a biomimetic approach to influence CM behaviors, strip, oval and wave-patterned mats were constructed by deposition of electrospun fibers on lithographic collectors, followed by precise stacking for cell co-cultures. CMs, CFs, and ECs were located on the patterned scaffolds with controlled cellular distribution in the respective regions and no across condition was found. Compared with those after strip and oval-patterned co-culture, CMs co-cultured on wave-patterned scaffolds displayed significantly greater cell viabilities, larger cell elongation ratios, stronger expressions of cardiac-specific Troponin I, connexin 43 and sarcomeric α-actinin and higher beating rates during 15 days of incubation. The responses of co-cultured CMs to quinidine, erythromycin and sotalol show good correlations with clinical observations in the beating rate and the prolongation of the contraction and relaxation time. The in vivo safety data reflected well with the concentrations for 50% of maximal effect (EC₅₀) after drug treatment on co-cultured CMs, which was determined from the changes in the corrected field potential duration (FPDc) against the drug concentrations. During 15 days of patterned co-culture, the interbeat intervals and fluctuations of the CMs indicated quick changes in response to haloperidol treatment and sufficient restoration of the original beating profiles after drug removal. This study demonstrates the capabilities of micropatterned co-culture of CMs to establish the cardiac function as a reproducible and reliable platform for screening cardiac side effects of drugs.