The lithium/fluorinated carbon (Li/CF_x) battery is a widely used solid-state lithium primary battery (LPB) with the highest ever known specific energy. However, the scarcity of lithium resources compels researchers to explore other types of batteries. Sodium batteries have been considered as the most suitable candidate power sources due to the low cost of sodium and its similar chemistry to lithium. In this work, we prepared two series of fluorinated carbon nanotubes (CNTs), which are fluorinated common CNTs (FCNT-x) and fluorinated graphitized CNTs (FGCNT-x). The effects of graphitization of the CNTs on the structure of fluorinated carbon nanotubes and C–F bonding are systematically studied. FGCNT-x prepared from the graphitized CNTs exhibit a lower ratio of F/C and weaker C–F bonding than those of the FCNT-x prepared from ordinary CNTs at the same fluorination temperature. FGCNT-0.81 exhibits very high specific capacities of 798.8 and 751 mA h g⁻¹ in lithium and sodium primary batteries, nearly reaching the theoretical capacity calculated using the F/C ratio. Besides, FGCNT-0.81 possesses a maximum energy density of 2006.6 W h kg⁻¹ in Li/CF_x batteries and 1733.4 W h kg⁻¹ in Na/CF_x batteries. The maximum power density for Li/CF_x and Na/CF_x batteries could reach up to 3861.1 (FCNT-0.81) and 1863.9 W h kg⁻¹ (FGCNT-0.81), respectively. It is proved that the binding energy of the C–F bond, the electronegativity and the polarization force of the metal ions collaboratively determine the electrochemical performance of the fluorinated CNTs. The new-type Na/CF_x primary batteries developed here are promising in low power applications such as high-energy density and low-cost batteries.