This study investigates the interplay of working memory, cognitive style, and behaviour. Year 8 (aged 13 years) students (n = 205) at a UK urban secondary school were tested to ascertain predictors of General Certificate of School Education (GCSE) achievement. Assessment included Riding's cognitive style dimensions, working memory capacity, and a profile of school behaviour. A stepwise multiple regression analysis indicated that behaviour, working memory, verbal-imagery style by working memory, wholist-analytic style by working memory, and verbal-imagery style by wholist-analytic style predicted GCSE outcome, accounting for 58% of the variance. Generally, poor GCSE grade points were predicted by low behaviour scores, and by low working memory capacity in analytics and verbalisers. The results are discussed in terms of possible interventions that may improve student GCSE performance. (Contains 6 tables.)