This study investigates how dynamic complexity influences the use of heuristics in a dynamic decision making task. 24 subjects performed a computerized task that required them to predict the dynamic change (t+1) of discrete interrelated variables (range 0-20). There were four experimental sessions that involved the co-evolution of eight variables over 40 time periods. The level of complexity of the dynamic change of each variable was determined by its distance to linearity. Five heuristics, characterized by different methods of effort-reduction (Shah & Oppenheimer, 2008), were formalized. Forecasting performance of participants was compared against each heuristic. The results revealed that subjects adapt their heuristics to the level of dynamic complexity. Our findings are discussed in relation to the conditions under which individuals adapt decision rules in complex dynamic tasks and on the consequences of such heuristic changes (Hogarth & Karelaia, 2007).