Wireless sensor networks have been used for a variety of purposes such as habitat monitoring, malicious target detection, and climate observation. Conventionally, stationary sensors are used to carry out sensing tasks in a region of interest. However, if the region is much larger than the sensing range of a single sensor, stationary sensor networks could incur many problems such as the complicated reciprocal effects between coverage and communication ranges. This paper proposes a method to collect sensing data by mobile sensors. The method first selects critical sensing locations using detection performance as a metric. The sensing tasks are assigned to a number of mobile sensors. These mobile sensors are arranged to collect sensing data through routes as short as possible. In particular, to balance the load, the routes should be planed in approximately equal in length. An approach, namely Balanced Route Planning (BRP), is developed to arrange mobile sensors' patrolling paths. The load balance condition of BRP is evaluated by simulations. The results show that BRP is a fair and effective scheme for allotting sensing tasks for mobile sensors.