We investigate the influence of stellar rotation on the H-alpha line formation in O-star winds. The 2-D wind model used is based on the kinematical approach by Bjorkman & Cassinelli (1993, BC), adapted to the parameter space considered in this paper. We discuss only those cases where the rotational rates are well below those that would induce an onset of disk formation. The influence of gravity darkening on the line formation is shown to be negligible, as long as appropriate averaged photospheric parameters (which then are a function of the rotational rate) are used. The distortion of the stellar radius from sphericity can likewise be neglected in most cases. Our investigations show that the H-alpha line formation is strongly affected by two processes which we call the resonance zone and the rho(2)-effect. The former process diminishes the emission near the line core and enhances the emission in both wings due to a twist in the resonance zones induced by differential rotation. The latter process leads to an increase in the overall emission due to the density contrast between the polar and the equatorial zones caused by the deflection of material towards the equator in the BC-model. We compare the line profiles from our 2-D models with those resulting from the conventional 1-D approach, as a function of absolute or projected rotational velocity, and inclination angle and mass-loss rate. It is shown that in all cases independent of inclination angle and rotational rate, the 1-D method - for a given mass-loss rate - yields the smallest wind emission. This in turn means that all mass-loss rates presently derived from H-alpha are overestimated, with typical errors of 20...30%. The maximum error introduced by this simplified approach is of the order 50...70% for O-Supergiants and occurs for sears with small nu(rot) sin i and observed nearly pole on. Moreover, our theoretical line shapes show a number of features actually found in the observations of rapidly rotating stars. Finally, the specific influence of the rotational rate and inclination angle which both, independently, modify the profiles in a distinctive way may provide us with a method for the determination of sin i from H-alpha line fits (in connection with the analysis of other spectral regions) in future investigations.