Many multitasking studies used randomized stimuli and stimulus onset asynchrony (SOA). Here we investigate whether and how sequence learning on position of stimuli and responses, as well as sequence learning on SOA support dual tasking. We combined two visual-manual Serial Reaction Time (SRT) tasks. Participants (N = 28 in Exp. 1 and N = 30 in Exp. 2) were first presented with a randomly sequenced two-choice task and afterwards the four-choice SRT task (followed a repeating sequence of length four). The SOA followed also a repeating sequence of length four. In test phases, either or both of the sequences were shifted to random. Results showed that sequence of position and delay were learned. Different from work on acquisition of interval sequences in single tasking, we found that the sequence of “what” and the sequence of “when” contributed independently to performance. Additionally, sequence knowledge can influence the task-order scheduling.