A reduced order kinetic model for NO (Nitric Oxide) prediction, based on the recently developed virtual chemistry methodology, is proposed in this article. Temperature and heat release are resolved through the optimization of a virtual main mechanism whereas NO emissions are reproduced through the optimization of a dedicated sub-mechanism. The proposed NO sub-mechanism is optimized over a learning database made of 1D-premixed and 1D opposed jet diusion flames. The approach is further assessed in a premixed 2D CFD laminar flame computation proposing a direct comparison against fully detailed chemistry.