This study aims to investigate the linear and nonlinear (non-compensation) effects of teacher leadership on teacher well-being dimensions through the three basic psychological needs dimensions (autonomy, competence, and relatedness). This quantitative cross-sectional study used partial least squares structural equation modelling and artificial neural network (ANN) for data analysis. Data were collected using survey questionnaires from 728 Malaysian primary school teachers. All the direct (linear) and indirect effects of teacher leadership on teacher well-being dimensions through teachers' basic psychological needs were significant. The ANN analysis revealed that competence, a dimension of teacher basic psychological needs, was the strongest predictor of workload well-being and student interaction well-being. Autonomy was the strongest predictor of organisational well-being. Implications and future studies are presented.