We study the problem of dynamic partitioning of the workload of yard cranes in a container terminal with the objective of expediting their service to the container-carrying vehicles, which in turn helps to reduce the vessel turnaround time. Because the workload distribution in the container storage yard changes dynamically, we propose to periodically divide container yard blocks into non-overlapping zones for cranes allocated to the yard blocks, then allocate each non-overlapping zone to a distinct crane. The non-overlapping zones are separated by a safety distance so the crane clash problem and the resulting delays in operations are eliminated. To obtain the optimal partitions for each planning period, we use real time data driven simulation to estimate the dynamic workload of each yard crane in the proposed yard crane workload partition algorithms. Our experimental results show that our partition scheme is indeed able to substantially reduce vehicle waiting time.