This paper constructs and structurally estimates a dynamic model of occupational choice where all occupations are characterized in a continuous multidimensional space of skill requirement using the data from the Dictionary of Occupational Titles and the NLSY79. This skill space approach allows the model to include hundreds of occupations at the three-digit census classification level without a large number of parameters. Thereby it provides more detailed analysis of occupations than previous papers. Parameter estimates indicate that skill demanding occupations offer higher returns to education and experience, which results in occupational sorting. They also suggest that the estimated skill prices by the OLS are severely biased due to this sorting.