Percutaneous closure is the treatment of choice for secundum-type atrial septal defects (ASD). Balloon sizing (BS) has been the method of choice for deciding on device size. Improved 2D- and 3D-transesophageal echocardiographic (TEE) imaging challenged the necessity of BS. Balloon sizing was performed with two additional techniques to measure the stretched dimension of the ASD. The 1st method uses a stiff guide wire which stretches the ASD and 2D TEE. The second technique uses 3D TEE. Two hundred and thirty-six patients with minimum 1-year follow-up were enrolled. The population was classified into three groups: BS (group 1) n = 90, 2D-TEE (group 2) n = 87, and 3D-TEE (group 3) n = 59. All groups showed a distinct correlation between the maximum baseline dimensions and the device size (R = 0.821). The relative expansion rate did not differ between BS and 3D-TEE. Group 2 (2D-TEE) showed a significantly lower expansion rate. Procedural success and complications did not differ statistically between the 3 groups. 2D TEE sizing was the simplest method without loss of accuracy. 3D sizing offers the advantage of accurate and fast shape assessment, but resulted in more undersizing. Accurate sizing of ASDs with a floppy septum remains a challenge.