Mounting evidence has highlighted the multifunctional characteristics of glutamine metabolism (GM) in cancer initiation, progression and therapeutic regimens. However, the overall role of GM in the tumour microenvironment (TME), clinical stratification and therapeutic efficacy in patients with ovarian cancer (OC) has not been fully elucidated. Here, three distinct GM clusters were identified and exhibited different prognostic values, biological functions and immune infiltration in TME. Subsequently, glutamine metabolism prognostic index (GMPI) was constructed as a new scoring model to quantify the GM subtypes and was verified as an independent predictor of OC. Patients with low‐GMPI exhibited favourable survival outcomes, lower enrichment of several oncogenic pathways, less immunosuppressive cell infiltration and better immunotherapy responses. Single‐cell sequencing analysis revealed a unique evolutionary trajectory of OC cells from high‐GMPI to low‐GMPI, and OC cells with different GMPI might communicate with distinct cell populations through ligand‐receptor interactions. Critically, the therapeutic efficacy of several drug candidates was validated based on patient‐derived organoids (PDOs). The proposed GMPI could serve as a reliable signature for predicting patient prognosis and contribute to optimising therapeutic strategies for OC.