Objective To explore a new method for atrial septal puncture in canine. Methods Ten healthy mongrel dogs, weighing 10-15 kg, were used in the present study. Femoral venous access was obtained after anesthesia; a 0.032 inch (1 inch=2.54 cm) guide wire and transseptal sheath were introduced into the superior vena cava(SVC).The guide wire was then withdrawn and the Brockenbrough needle was advanced to the tip of the catheter(1 cm to the tip, pointing the 12 o'clock direction). Under the posteroanterior view, the needle and sheath assembly were gradually withdrawn from the SVC and gently rotated clockwise into the right atrium. When reaching the lower third of the RA, the tip of catheter was slipped to the left. The entire transseptal apparatus was adjusted to 3 o'clock direction to allow the tip pointing the landmark (location of the second rib above the heart shadow) in the right anterior oblique 40°-50° view. Gently push the sheath till resistance, and then the needle was manipulated to the fossa ovalis (FO) under fluoroscopy. After the location of the needle tip in the left atrium (LA) was confirmed by angiogram, the catheter was advanced over the needle to the LA. After removal of the needle, the 0.032 inch LA guide wire was introduced into the LA. The sheath assembly was advanced into LA over the guide wire. After removal of the guide wire and the dilator of sheath, a heparin bolus of 1 800-3 000 (80-100 units/kg) was administered. Electrocardiogram(ECG) and transthoracic echocardiography(TEE) were performed 30 min after transseptal puncture. Five dogs were sacrificed after procedure to observe the location of the puncture, and other dogs were followed up for 1 month. Results The first attempt of puncture at TS crossing was successful in 8 (80%) dogs, and the second attempt was successful in 2 (20%) dogs. One dog died due to pericardial tamponade. ECG and TEE showed no abnormal findings in other dogs. In the five sacrificed dogs no bloody pericardial effusion was found, and the sites of TS puncture were at the center of FO. There were no other complications in the other dogs. The fluoroscopy time was (9.2±6.4) min and the procedure time was (45±12.6) min in our study. Conclusion Our new method is safe, simple and effective to perform TS puncture under fluoroscopic guidance from fixed bony landmarks.