This paper presents a technique to detect and track multiple humans in a video sequence. Human detection is done using Fast Mean Shift approach. Multiple humans in a video sequence are differentiated on the basis of their color and appearance. The occlusion issue is also addressed in the paper and is removed using the particle filtering and appearance matching techniques. The technique presented for detecting and tracking are efficient and computationally inexpensive. Experimental work shows that in case of low image quality and occlusion, this approach works properly.