The present study examined age differences in time-based prospective memory (TBPM) in primary school age children and tested the role of self-initiated memory retrieval and strategic time monitoring (TM) as possible developmental mechanisms. Fifty-four children were recruited from local primary schools (27 younger children, mean age = 7.2 ± 0.55 years, and 27 older children, mean age = 9.61 ± 0.71 years). The task was a driving game scenario in which children had to drive a vehicle (ongoing task) and to remember to refuel before the vehicle runs out of gas (TBPM task, i.e., the fuel gauge served as child-appropriate time equivalent). Fuel gauge was either displayed permanently (low level of self-initiation) or could only be viewed on demand by hitting a button (high level of self-initiation). The results revealed age-dependent TBPM differences with better performance in older children. In contrast, level of self-initiated memory retrieval did not affect TBPM performance. However, strategies of TM influenced TBPM, as more frequent time checking was related to better performance. Patterns of time checking frequency differed according to children's age and course of the game, suggesting difficulties in maintaining initial strategic TM in younger children. Taken together, the study revealed ongoing development of TBPM across primary school age. Observed age differences seemed to be associated with the ability to maintain strategic monitoring.