Gene Electro-Transfer (GET) is a powerful method of DNA delivery with great potential for medical applications. Although GET has been extensively studied in vitro and in vivo, the optimal parameters remain controversial. 2D cell cultures have been widely used to investigate GET protocols, but have intrinsic limitations, whereas 3D cultures may represent a more reliable model thanks to the capacity of reproducing the tumor architecture. Here we applied two GET protocols, using a plate or linear electrode, on 3D-cultured HCC1954 and MDA-MB231 breast cancer cell lines grown on a novel collagen-free 3D scaffold and compared results with conventional 2D cultures. To evaluate the electrotransfer efficiency, we used the plasmid pEGFP-C3 encoding the enhanced green fluorescent protein (EGFP) reporter gene. The novel 3D scaffold promoted extracellular matrix deposition, which particularly influences cell behavior in both in vitro cell cultures and in vivo tumor tissue. While the transfection efficiency was similar in the 2D-cultures, we observed significant differences in the 3D-model. The transfection efficiency in the 3D vs 2D model was 44% versus 15% (p < 0.01) and 24% versus 17% (p < 0.01) in HCC1954 and MDA-MB231 cell cultures, respectively. These findings suggest that the novel 3D scaffold allows reproducing, at least partially, the peculiar morphology of the original tumor tissues, thus allowing us to detect meaningful differences between the two cell lines. Following GET with plate electrodes, cell viability was higher in 3D-cultured HCC1954 (66%) and MDA-MB231 (96%) cell lines compared to their 2D counterpart (53% and 63%, respectively, p < 0.001). Based on these results, we propose the novel 3D scaffold as a reliable support for the preparation of cell cultures in GET studies. It may increase the reliability of in vitro assays and allow the optimization of GET parameters of in vivo protocols.