Complex problem solving (CPS) as a cross-curricular competence has recently attracted more attention in educational psychology as indicated by its implementation in international educational large-scale assessments such as the Programme for International Student Assessment. However, research on the development of CPS is scarce, and the few existing studies are cross-sectional. Therefore, the present study analyzed CPS development with longitudinal data on adolescent students collected over a period of 2 years. CPS development was estimated with latent growth curve models, and fluid reasoning, age, and sex served as predictors. CPS growth patterns were positive and linear and were positively related to fluid reasoning. Older students performed better on initial CPS but exhibited smaller increases (i.e., less development) in CPS performance. No meaningful sex differences in initial CPS or CPS development were found except that boys showed slightly better initial performance on the CPS dimension knowledge application. These results present an important first step in the investigation of CPS development.