We used data from a large, longitudinal study of children in the community, the NICHD Study of Early Child Care and Youth Development, to examine how well earlier measures of delay capacity, inhibitory control, planning, and attention predicted symptoms of attention deficit hyperactivity disorder (ADHD) assessed in third grade. Children with elevated symptoms of both inattention and hyperactivity-impulsivity (n = 57) and with inattentive symptoms only (n = 80) were identified via mother and teacher reports using the "or" rule, as were children without significant symptoms (n = 790). Multinomial logistic regression analyses indicated that poorer performance on earlier measures of resistance to temptation, delay of gratification, response inhibition, attention, and planning obtained from 36 months to 1st grade predicted membership in the two symptom groups relative to the comparison group in 3rd grade, albeit with somewhat different patterns of predictors. Controls for 36 month school readiness and externalizing symptoms indicated that these results were generally robust and not an artifact of initial cognitive or behavioral differences. Implications for developmental models of ADHD are discussed.