In the present work, we studied the near-UV photoinduced decarboxylation of o-acetylphenylacetic acid with complete active space self-consistent field and density functional theory. It was found that irradiation at approximately 300 nm results in o-acetylphenylacetic acid in the S(1)((1)npi*) state, which is followed by a rapid relaxation and efficient intersystem crossing to the T(1)((3)npi*) state via the S(1)/T(2)/T(1) three-surface intersection. The 1,5-H shift has a barrier of 1.9 kcal x mol(-1) on the T(1) pathway to the triplet 1,4-biradical, which is in good agreement with a rate constant of about 10(10) s(-1) inferred experimentally for o-alkylphenyl ketones. The subsequent reactions occur with little probability from the triplet 1,4-biradical, due to relatively high barrier or high endothermicity for the spin-conservation triplet pathways. As a result, intersystem crossing to the lowest singlet state takes place prior to the subsequent reactions. Four isomers, (3)BRi (i = 1-4), were found to be stable for the triplet 1,4-biradical. The calculated energy gap indicates that the (3)BR3/(3)BR2 ratio is close to 1:1, and populations of (3)BR1 and (3)BR4 are less than 1% at thermal equilibrium. Like the triplet 1,4-biradical, four stable isomers of (1)Xi (i = 1-4) were determined in the lowest singlet state. Because of the relatively high barrier (approximately 30 kcal x mol(-1)) on the isomerization pathways, the thermal equilibrium is not established among the four singlet isomers, which is different from the situation for the triplet 1,4-biradical. In this case, the subsequent reactions proceed mainly from the (1)X2 and (1)X3 isomers that correspond to (3)BR2 and (3)BR3. There is only one predominant pathway from the (1)X2 isomer, namely, the reversed-H shift to the initial reactant of o-acetylphenylacetic acid. However, several possible pathways exist for the (1)X3 deactivation: intramolecular cyclization, unimolecular decarboxylation, and the parent acid-catalyzed bimolecular decarboxylation. The unimolecular decarboxylation is not in competition with the cyclization. But the cyclization reaction is prevented by the parent acid-catalyzed bimolecular decarboxylation, which is responsible for the products of CO(2) and o-acyltoluene observed experimentally. The (1)X3/(1)X2 ratio is nearly equal to that for (3)BR3/(3)BR2, which indicates that the decarboxylation reaction has a quantum yield close to 0.5. The o-acyltoluene product was isolated, and its yield was experimentally estimated to be in the 50% range.