The aerobic oxidation of alkenes to carbonyls is an important and challenging transformation in synthesis. Recently, a new P450-based enzyme (aMOx) has been evolved in the laboratory to directly oxidize styrenes to their corresponding aldehydes with high activity and selectivity. The enzyme utilizes a heme-based, high-valent iron-oxo species as a catalytic oxidant that normally epoxidizes alkenes, similar to other catalysts. How the evolved aMOx enzyme suppresses the commonly preferred epoxidation and catalyzes direct carbonyl formation is currently not well understood. Here, we combine computational modelling together with mechanistic experiments to study the reaction mechanism and unravel the molecular basis behind the selectivity achieved by aMOx. Our results describe that although both pathways are energetically accessible diverging from a common covalent radical intermediate, intrinsic dynamic ef fects determine the strong preference for epoxidation. We discovered that aMOx overrides these intrinsic preferences by controlling the accessible conformations of the covalent radical intermediate. This disfavors epoxidation and facilitates the formation of a carbocation intermediate that generates the aldehyde product through a fast 1,2hydride migration. Electrostatic preorganization of the enzyme active site also contributes to the stabilization of the carbocation intermediate. Computations predicted that the hydride migration is stereoselective due to the enzymatic conformational control over the intermediate species. These predictions were corroborated by experiments using deuterated styrene substrates, which proved that the hydride migration is cis-and enantioselective. Our results demonstrate that directed evolution tailored a highly specific active site that imposes strong steric control over key fleeting biocatalytic intermediates, which is essential for accessing the carbonyl forming pathway and preventing competing epoxidation.