Vehicles can provide various sensing abilities and unlimited communication capabilities and are taken as an important platform for collecting sensory data for multiple on-going urban sensing tasks. However, new challenge arises for selecting vehicles with different incentive requirements, various sensing abilities and uncontrollable mobilities to best satisfy heteroid sensory data requirements of multiple concurrent applications under budget constraints, but with sparsely research exposure. This paper proposes a multi-task-oriented participant selection strategy to tackle the above mentioned challenge. The difference between data requirements of multiple tasks and data collection expectation of a set of vehicles are converted to a multi-aim optimization problem, and a greedy-algorithm-based participant selection strategy is designed to solve it. Real dataset based simulation show that under the same incentive costs condition, the proposed participant selection strategy can obtain more comprehensive sensory data than selecting vehicles randomly.