A massive amount of video data is recorded daily for forensic post analysis and computer vision applications that are used for the analysis of this video data, often perform Multiple Object Tracking (MOT). Advancements in image analysis algorithms and global optimization techniques have improved the accuracy of MOT, often at the cost of slow processing speed which limits its applications to small video datasets. With a focus on fast MOT, this thesis introduces a greedy data association technique (GDA) for MOT, which finds a locally optimum solution with a low computational overhead. To further enhance the computational speed of MOT for large video datasets, three MapReduce-based parallel techniques are introduced. The performance analysis using a set of benchmark video datasets with system prototypes shows that the proposed techniques are significantly faster than the existing state-of-the-art methods.