Aromatic hydroxy acids, the compounds of large industrial importance, can be prepared in the Kolbe-Schmitt reaction, i.e. a carboxylation reaction of alkali metal phenoxides (MOPh) and naphthoxides (MONaph). On the basis of the experimental results two contradictory reaction mechanisms have been proposed: the one of direct carboxylation, and the other involving initial formation of the MOPh-CO2 or MONaph-CO2 complex. Previous theoretical investigations of the carboxylation reaction of sodium 2-naphthoxide, performed by means of the B3LYP method, confirmed the initial formation of the NaONaph-CO2 complex, and showed that the carbon of the CO2 moiety performs an electrophilic attack at C1 of the ring, leading to the formation of sodium 2-hydroxy-1-naphthoate (E1). Surprisingly, transition states for possible electrophilic attacks at C3 and C6 were not revealed, and the formation of other two products (E3 and E6) was explained by a number of consecutive rearrangements. In addition, this mechanism includes a reaction step with rather high activation energy. Since more sophisticated functionals are today available, the aim of this work is to reinvestigate the mechanism of the Kolbe-Schmitt reaction of NaONaph in all three positions (1, 3, and 6). Our investigations with the M062X method demonstrated that CO2 and NaONaph can spontaneously build two complexes: B (the one previously reported) and C. While B cannot be further transformed to yield the reaction products, the CO2 moiety in C takes perfect position for electrophilic attacks at all three sites of the ring. These attacks are realized via the transition states TS1, which lead to the formation of the new C-C bonds, and corresponding intermediates D. In the next, bimolecular reaction step two D intermediates exchange the protons adjacent to the CO2 groups. These intermolecular reaction steps require significantly lower activation energies in comparison to the intramolecular proton shift from C to O. The carboxylation reaction in the position 6 is both kinetically and thermodynamically unfavourable, whereas the pathways in the positions 1 and 3 are competitive. Pathway 1 requires the lowest activation energies, but E3 is significantly more stable than other two products. In accord with these findings are the experimental results which show that, at very low temperature (293 K) only E1 is formed at low yield, whereas the yields of E3 and E6 increase with the increasing temperature. Since the Kolbe-Schmitt reaction is experimentally performed at relatively high temperatures (around 500 K), the main product is thermodynamically most stable E6.