I am solving a flow over an airfoil using unstructured grids(segregated), I would like to whether decreasing the momentum to 0.7 will affect the accuracy of the results, and also the other factors.

no. reducing urf's wont affect result but will make it easier (more robust and less likely to diverge) but it will be slower to converge.

Armin,

View under-relaxation as de-acceleration. It will help prevent bumpy rides. However you will finally reach the destination but with more number of iterations. It only limits large changes in a particular variable every iteration.

It would be useful to create a monitoring point/surface. You can monitor pressure of drag coefficient on the airfoil. Ones the quantity monitored stops changing with iteration you can stop your simulation. remember to monitor the peak value instead of the average value.

Regards, Ant

thank you Ant,I am using unstructured grid for solving Euler equations, interestingly by reducing the momentum factor, the number of iterations reduces and the solution converges faster, Could you please tell me the reason

There is no reason for the convergence rate to be monotonic with the relaxation factor. Usually there is an optimal value, but it is case specific and hardly ever predictable. Therefore former experience and/or trial-and-error are needed to enhance the convergence.