This paper offers a methodology based on the concept of time-space prism in choice set building for destination choice models.  Using datasets that are widely available and already used by planning agencies, a representation of available activity opportunities at the US Census block level is generated, and a dynamically changing environment is virtually represented to account for changing travel time and available opportunities during a day.  As a part of the methodology, an algorithm for computing potential path area (PPA) based on spatio-temporal constraints is offered.  Using the algorithm PPAs for different settings for different times in a day and individual temporal decisions are computed, and several comparisons are conducted for the PPAs.  The settings that we identified to be important in choice set building are size of time budget, minimum activity duration, daily profile of available opportunities, travel times varying by time of day, and temporal flexibility in scheduling activities.  Their influence as implications on choice set building and amount of opportunities available is explained using illustration, and the results confirm that not only the choice set itself but also the characteristics of the destination alternatives change according to the factors that we identified.