The issues of accessibility, management, storage and organization of Learning Objects (LOs) in education systems are a high priority of the Thai Government. Incorporating personalized learning or learning styles in a learning object management system to improve the accessibility of LOs has been addressed continuously in the Thai education system. A proposed Learning Object Management Model (LOMM) is discussed in this paper which aims to adapt and optimize the learning process based on characteristics of the individual learners. This study aims to find the correlation between learning styles and LOs characteristics in the LOMM. Decision Tree and Apriori algorithms were used to generate a predictive model for the classification of learners. Development of the predictive model was based on survey results from 1,586 high school students in Nakhon Ratchasima province, Thailand. The diverse LOs characteristics were analyzed in order to find the correlation with learning styles of the learners. The classification model consists of 24 sub-models used to predict a learner's class based on 8 groups of LOs characteristics. The best accuracy obtained in the study was 80.23%. Finally, for the next phase this approach has been designed to support the proposed LOMM and it is expected that it could be readily applied to other e-learning systems and digital repositories.