Background: Suicide is the second leading cause of death among youth and young adults. Improving the understanding of social-ecological factors associated with youth suicide behaviors may highlight important ideas for prevention. Objective: From a social-ecological theoretical framework of suicide, this study aims to test whether parental monitoring and school academic engagement predict suicidal behavior and substance use in youth, and by gender. Methods: Data comes from a large cross-sectional U.S. study--the 2017 National Survey on Drug Use and Health (NSDUH)--which includes U.S. adolescents (N = 12,884) aged 12-17 years old. A structural equation model (SEM) and the MIMIC (Multiple Indicators and Multiple Causes) model of SEM were used to test the proposed conceptual model using AMOS Version 25. Results: The final structural model had an adequate goodness of fit [X[superscript 2](df = 84, N = 12,884) was 2352.58, p < 0.001]. The other fit statistics indicated good fit with RMSEA = 0.044 and CFI = 0.94. Fifty-eight percent of the variance in suicide behaviors was accounted for by the model. Conclusions: These findings underscore critical associations between suicidal behavior and relational (parental monitoring), community (school academic engagement) and individual (substance use) social ecological factors. Gender paths in the model did not differ significantly. These factors suggest important prevention targets for reducing suicide risk among adolescents.