As mentioned previously, Mean-shift is probably not the best tracker ever. But keep in mind that this is nevertheless a fast visual tracker which makes it very suitable for real-time applications.

The solution proposed before by Mr. Dey, is actually quite interesting but only if you consider a static camera (a wide literature exists on those background subtraction approaches ...).

First, thank you very much for your answers Mr Dey and Mr Rameau, exactly in my first contribution I worked with a static camera but now I want to work with a moving camera, and after searching I found that the Mean Shift algorithm is very used to track object especially in real-time applications.I am very grateful if you can help me ..

Meanshift is the most basic algorithm for object tracking. For practical using, I think Camshift, which stems from meanshift, may be a better choice  for tracking moving objects. 

Mean-shift is not the only/best solution...depend upon the statement and conditions of the problem. you may consider Kalman filter for linear motion and extended kalman filter or particle filter for non-linear scenarios... if u have enough data then Neural Networks will give you good results ever.

Please have a look at the survey paper linked to below on tracking techniques. There are very many algorithms for tracking in radar, sonar and video systems. The most basic of these is the nearest-neighbour filter (based on alpha-beta or Kalman filters). For video you typically track the image centroid (which is like the peak of the object PDF used in the mean shift algorithm), but there are also techniques that try to track extended objects (more than 1 pixel). When you say "track people", it depends very much on the resolution of the imaging system, e.g. short or long range. There will also be problems with occlusion/shadowing, etc. Either way, in its most basic form, this sounds like a multiple target tracking problem.

Article Taxonomy of multiple target tracking methods

Hello Khadija,

We have recently published an efficient and effective people tracking method for video-surveillance in shopping malls. Take a look to our publication for more information and results, perhaps it can help you:

http://dx.doi.org/10.1016/j.eswa.2015.06.016

Good luck with your research!

Article Expert Video-Surveillance System for Real-Time Detection of ...

Thank you Dr. Roberto Arroyo for your help.

If you are working on pedestrian detection and tracking you may consider HOG algorithm,..

yes, one of my student used it to track football player