When Influence Maximization (IM) is applied to social network to maximise the network coverage, it becomes an effective mechanism for marketing applications. In this paper, we focus on a specific influence maximization problem, i.e., selecting a set of seeds on twitter to maximise information propagation, which can be used for information reaching out in marketing campaigns. The proposed approach is taking into the consideration of social ties, user interactions, and information propagation on Twitter. The influence probability is calculated according to users' action history including tweet, favourite, mention/reply, and retweet. An information diffusion model is proposed with the capability to simulate the dynamic process of information spread on Twitter. A concise heuristic algorithm is developed for influence maximization accordingly. Experimental results and analysis are provided based on a real Twitter network including 3,292 users in Darwin city in Australia.