Scheduling with periodic maintenance has been widely studied. However, multi-agent scheduling with simultaneous considerations of periodic maintenance has hardly been considered until now. In view of this, this research focuses on the problem of scheduling jobs that come from two agents on a single machine under periodic maintenance constraint with the objective of minimising the total completion time of the jobs of the first agent while keeping the maximum tardiness of other agent below or at a fixed level UB. We present some new dominance properties for this strongly NP-hard problem. And next, using these properties, we develop a novel imperialist competitive algorithm for the problem. Various parameters of the proposed algorithm are reviewed by means of Taguchi experimental design. For the evaluation of the proposed ICA, problem data was generated to compare it against a genetic algorithm. The results of computational experiments show the good performance of the proposed algorithm.