Puncture of the temporomandibular joint (TMJ) is a minimally invasive treatment for various jaw disorders. This study used a cadaveric model to evaluate the procedure from two entrance points with respect to hit ratio and possible complications, such as extraarticular extrusion of injection fluid. Ten heads, embalmed with Thiel's method, were investigated. A straight line drawn with a colored pen connected the center of the tragus and the lateral canthus. The first portal “A” was located at a distance of 1 cm anterior and 2 mm caudal from the center of the tragus. Portal “B” was located 2 cm anterior and 1 cm caudal starting from the same reference point. Punctures “A” and “B” were performed alternately on the right and left sides. Specimens were dissected and the local distribution of the injected latex was recorded. With Approach A, four punctures (40%; 4/10) reached the TMJ, whereas with Approach B, six injections (60%; 6/10) entered the TMJ. There were no statistically significant differences between the tested puncture methods (P = 0.0317) and body sides (P = 1). With each method, for example, 35% (7/20) each, the injected latex was either periarticular or retromandibular. In a further 20% (4/20), it was located subperiosteally alongside the ramus of mandible. The latex was injected into the infratemporal fossa and the external acoustic meatus in one case each (each 5%). There was no statistically significant difference between the techniques. The adjacent anatomy has to be kept in mind during TMJ puncture as the complication rate was remarkably high, suggesting that ultrasound guided intraarticular injection could improve the hit rate. Clin. Anat., 33:683–688, 2020. © 2019 Wiley Periodicals, Inc.