Where traffic is congested and all the curb spaces are occupied, some drivers are probably cruising for parking. Cities can eliminate this cruising by charging demand-based prices for curb parking to ensure one or two open spaces on every block. Drivers will then usually find convenient places to park within a short walk of their destinations. But if cities charge demand-based prices for curb parking, how will drivers choose where to park, and what will they pay? To answer these questions, I have examined how four variables-parking duration, number of persons in the car, walking speed, and value of saving time spent walking-determine parking choices when prices increase as drivers approach their destinations. Short-term parkers, carpools, slow walkers, and drivers with a high value of saving time will park closer to their destinations. Long-term parkers, solo drivers, fast walkers, and drivers with a low value of saving time will park farther away. This spontaneous, self-organizing pattern of parking choices responding to demandbased parking prices will minimize the collective cost of the time drivers spend walking to and from their destinations. Demand-priced curb parking will also reduce uncertainty about travel times. Drivers delayed by traffic congestion can save time at the end of their trips by paying a higher price to park closer to their destinations. Demand-priced curb parking at the end of a trip can serve as a buffer allowing late-arriving drivers to buy time when they need it most.