The purpose of this paper is to present the data analysis obtained from our interview study, which showed that participants had different individual mobile learning (hereafter, abbreviated as m-learning) preferences. The understanding of these preferences for different m-learning requirements can be used as a foundation for building successful personalised m-learning applications catered to learners' individual m-learning needs. Participants' dynamic m-learning preferences (including location of study, noise/distraction level in a location and time of day) are described. We propose a context-aware personalised m-learning application based on these m-learning preferences. Six scenarios are given to illustrate the m-learning preferences of different learners. The system architecture consists of a learner profile, personalisation mechanism and learning object repository. An initial m-learning preference questionnaire is used to obtain learners' dynamic m-learning preferences. Current context values are retrieved from context-aware technologies. Appropriate learning objects are selected to learners based on their preferences and context values.