Vision guided robotic operation has been widely used in industrial applications. Especially, visual servoing has a glory prospect use in this scenario. It's widely known that the eye-in-hand configuration in visual servoing architectures is categorized as a highly coupled and nonlinear system. With the decoupling properties, it is very suitable to use some features deduced from moments to control the whole six Degrees of Freedom (DOFs) of a camera. In this paper, some new moment based features are given to control the rotational motion around different axis and a simple adaptive factor controller is introduced to further enhance the performance of the servo system. As a result, we get a greater convergence domain and better dynamic properties.