Background Few studies have investigated head impacts in youth ice hockey, none of which have reported impact mechanisms. Objective To investigate head impact characteristics in youth ice hockey. Design Video analysis. Setting 2013/2014 Calgary bantam (13–14 years) ice hockey season. Methods A previously compiled video database of 7260 bantam ice hockey player-to-player contacts from 22 games was searched for head impact cases. Eight games were randomly selected, two elite and six non-elite, from which head impact cases were analysed. Results A total of 254 head impact cases were identified, which represented 3.5% of all player-to-player contacts at a rate of 11.5 head contacts per game. A total of 100 head impact cases were analysed. Two-thirds of all cases (67%) occurred in close proximity to the boards and 11% of all cases resulted in a penalty. Over half of all impacts (55%) were to the side of the helmet, followed by the cage (29%), rear (7%), front (6%) and top (2%). The primary impacting object was an opposing player in 69% of all cases with the most common being the shoulder (31%), helmet (12%) and glove (10%). The impacting object was the glass and boards for 17% and 11% of all cases, respectively. A secondary impact occurred in 21% of all cases, which was most commonly to the side of the helmet and impacting the glass. One case involving a tertiary impact was identified, which comprised of two impacts to the shoulder of an opposing player and then an impact against the boards during the subsequent fall. Conclusions Impacts in youth ice hockey games are typically to the side and cage of helmets by an opposing player. Helmet performance and standards testing should include representative impacts by compliant surfaces to simulate player-to-player contact.