Data collection is one of the most important task in wireless sensor networks where a set of sensor nodes measure properties of a phenomenon of interest and send their data over a routing tree to the sink. In this paper, we propose a new simulated annealing based tree construction algorithm (SATC) to aggregate data with a collision-free schedule for node transmissions. SATC minimizes the time duration of delivering aggregated data to the sink. In the proposed algorithm, the average time latency is considered as the fitness function for the SA algorithm and it is evaluated based on Routing aware MAC scheduling methods. The efficiency of the proposed algorithm is studied through simulation and compared with existing state-of-the-art approaches in terms of average latency and average normalized latency.