Affective temperaments are inherited parts of personality determining mood and activity, affecting the management of somatic conditions. We aimed to investigate the association between affective temperaments, depressive symptoms, and self-care (physical activity, smoking, alcohol consumption), and their effect on glycemic control, among patients with type 2 diabetes (T2DM) in general practice, in a cross-sectional study enrolling 338 consecutive patients from six primary care practices in Hungary. A self-administered questionnaire (history, anthropometric, socioeconomic, laboratory parameters), the Beck Depression Inventory (BDI), the Hamilton Anxiety Scale, and the Temperament Evaluation of Memphis, Pisa, Paris, and San Diego Autoquestionnaire were used. Cyclothymic affective temperament determined HbA1c levels in regression analysis (p = 0.002), and the BDI score (p = 0.048). In causal mediation analyses, cyclothymic affective temperament was directly associated with higher HbA1c (p = 0.008). Hyperthymic affective temperament was indirectly associated with lower HbA1c, mediated by BDI (p = 0.034). Depressive, anxious, and irritable affective temperaments, and lifestyle factors were not associated with HbA1c neither in regression nor in mediation analysis as direct or mediating factors. Among primary care patients with T2DM, cyclothymic temperament correlates with worse glycemic control, independently of depressive symptoms. Hyperthymic temperament reduces depressive symptoms, thereby improving glycemic control. Identifying affective temperaments may improve diabetes care.