Cognitive theories suggest that biased suicide-related attentional processing confers risk for suicide ideation (SI) and attempts (SAs), but studies to date, which have been mixed, have not compared lower- and higher-order measures of attentional processes. The present study examined attention fixation, attention disengagement, and semantic interference as predictors of future SI and SA among adolescents, a high-risk age group for onset of SI and SAs. Adolescents (N = 135) (75 % female), ages 12-19 (M = 15.4, SD = 1.9), who presented to emergency or outpatient departments with SI (n = 84) or a SA (n = 51) completed an attention disengagement task (ADT), Suicide Stroop task, and the Attentional Fixation on Suicide Experiences Questionnaire (AFSEQ). A subsample of adolescents completed 3-month (n = 98) and 12-month follow-up assessments (n = 96) of SI severity and SA (n = 118) (based on adolescent or caregiver report). AFSEQ score predicted SI severity at 3- and 12-months, respectively, b = 0.47, p <
.05
b = 0.44, p <
.01, adjusting for ADT and relevant covariates, and also predicted SA at follow up, OR = 1.06, 95 % CI = 1.00-1.12, p <
.05, adjusting for SA history, but not after adjusting for SI. Greater suicide-related attention disengagement difficulty predicted greater SI severity at 12 months, b = 0.03, p <
.05, but not at 3 months. Self-reported, higher-order, attentional fixation may better predict future suicide-related risk among diverse adolescents than lower-order cognitive tasks assessing suicide-related attention disengagement or semantic interference.