Collisions with fixed obstacles during alpine ski competitions or recreational skiing seldom occur, but if they do the consequences can be severe, particularly in the case of competitions in which safety barriers are mandatory. Despite its importance, a standard method for evaluating the crashworthiness of such equipment has not yet been established. The aim of this work was to develop a full scale impact test method suitable for safety barriers commonly installed during Ski World Cup events. The construction and layout of different ski safety barriers, the type of rear support, and the impact speed varied throughout the tests. Peak decelerations and penetration values were measured during the impact of a dummy against a barrier. Tests were performed on B-nets with different poles and on air mattresses and foam mattresses placed on ground soil with properties similar to average snow. The mattresses were supported either by poles or by a rigid concrete wall. Tests on foam mattresses placed against the wall were conducted at increasing speeds in order to evaluate the relationship between peak deceleration and impact speed. A tower pendulum 30 m in length was used to accelerate a wooden dummy with a mass of 85 kg up to a maximum speed of about 65 km/h. The dummy was instrumented at the center of mass with a triaxial accelerometer, equipped with a data acquisition system, and released just before impact. Impacts were filmed using a high speed video camera. Fiberglass poles were the most effective in containing the dummy. Tests on foam mattresses against the concrete wall revealed a linear relationship between peak deceleration and speed. The method allowed analysis of the behavior of mattresses and B-nets, giving quantitative indications to manufacturers and to ski safety operators for the appropriate selection of safety barriers during competitions and in ski resorts.