Hemagglutinin (HA)-inhibiting antibodies contribute to the immune defense against influenza infection. However, there are insufficient data on the extent of correlation between vaccine-elicited HA antibodies and protection in children against different influenza strains, particularly when comparing live attenuated influenza vaccines (LAIV) versus inactivated influenza vaccines (IIV). We measured postvaccination hemagglutination-inhibition (HAI) titers in 3-15-year-old participants of a cluster-randomized controlled trial of trivalent LAIV(3) versus IIV(3) in Canadian Hutterite colonies. We assessed HAI titers as predictors of symptomatic, reverse transcription polymerase chain reaction (RT-PCR)-confirmed influenza over 3 influenza seasons using Cox proportional hazards regression models with vaccine type as a covariate. For each log2 unit increase in postvaccination HAI against A/H1N1 in 2013-2014, A/H3N2 2014-2015, and B/Yamagata in 2013-2014 (each the predominant circulating strain for the respective influenza season), the reduction in the risk of confirmed influenza was equal to 29.6% (95% confidence interval [CI], 17.1%-39.5%), 34.8% (95% CI, 17.2%-47.9%), and 31.8% (95% CI, 23.8%-38.5%), respectively. No reduction in the risk of influenza was observed with B/Yamagata-specific HAI titers in 2012-2013, which was dominated by a mixture of Yamagata and Victoria strains. Despite the overall lower HAI titers in the LAIV3 group, both H1N1 and H3N2 HAI titers were associated with protection against subtype matched influenza. Both LAIV3- and IIV3-elicited HA antibodies are associated with protection against influenza infection in seasons when the vaccine strains match the circulating influenza strain subtypes, supporting the use of HAI as a correlate of protection for both vaccine types in children.