We investigate the online bandit learning of the monotone multi-linear DR-submodular functions, designing the algorithm 𝙱𝚊𝚗𝚍𝚒𝚝𝙼𝙻𝚂𝙼 that attains O(T^2/3log T) of (1-1/e)-regret. Then we reduce submodular bandit with partition matroid constraint and bandit sequential monotone maximization to the online bandit learning of the monotone multi-linear DR-submodular functions, attaining O(T^2/3log T) of (1-1/e)-regret in both problems, which improve the existing results. To the best of our kno…