Despite the crucial role of aromatic-derived secondary organic aerosol (SOA) in deteriorating air quality, its formation mechanism is not well understood, and the dependence of aromatic SOA formation on nitrogen oxides (NOx) is not captured fully by most SOA formation models. In this study, NOx-dependent mechanisms of toluene and m-xylene SOA formation are developed using the gas-phase Caltech Atmospheric Chemistry Mechanism (CACM) coupled to a gas/aerosol partitioning model. The updated models were optimized by comparison to eighteen chamber experiments performed under both high- and low-NOx conditions at the University of California – Riverside. Correction factors for vapor pressures imply uncharacterized association chemistry, likely in the aerosol phase. The newly developed model can predict strong decreases of m-xylene SOA yield with increasing NOx. Simulated SOA speciation implies the importance of ring-opening products in governing SOA formation (up to ~40-60% for both aromatics). Speciation distributions under varied NOx levels imply that competition between hydroperoxide radical and NO for reaction with a bicyclic peroxide radical may not be the only factor influencing SOA formation. The reaction of aromatic peroxy radicals with NO competing with self-cyclization also affects NOx-dependence of SOA formation. Comparison of SOA formation yield and composition between toluene and m-xylene suggests aldehyde/ketone chemistry from a ring-opening route and a phenolic route play important roles in governing SOA formation, with their relative importance likely varying due to the number of methyl groups on the aromatic ring. Sensitivity studies of the NOx/Aromatic/SOA system with a range of NOx and hydrocarbon concentrations are carried out for toluene, m-xylene, and a mix of toluene and m-xylene. The impact of a non-aromatic hydrocarbon not expected to form SOA (propene) also is investigated. The effect of temperature variations on model predictions of the NOx effect on SOA formation is also explored. The updated box model will be implemented into a three-dimensional model to evaluate NOx effect on aromatic SOA formation on a larger scale.