This paper presents the experimental results of a study investigating the impact of machining parameters on 3D surface roughness after rough, dry milling. The following 3D roughness parameters were analysed: Sa (arithmetic mean height), Sq (root mean square height), Sz (maximum height), Sku (kurtosis), Ssk (skewness), Sp (maximum peak height), and Sv (maximum pit height). Roughness measurements were made on the end face of the specimens. Additionally, 3D surface topography maps and Abbot-Firestone material ratio curves were generated. Carbide end mills with variable rake and helix angles were used in the study. Experiments were conducted on AZ31B magnesium alloy specimens using a contact-type profilometer. The machining process was conducted using the parameters of so-called high-speed machining. Three variable technological parameters were analysed: cutting speed vc, feed per tooth fz, and axial depth of cut ap. The results showed that the surface roughness of the rough-milled specimens depended to a great extent on the tool geometry and applied machining parameters. Feed per tooth was found to have the greatest impact on surface roughness parameters. Lower values of the analysed surface roughness parameters (and therefore higher surface quality) were obtained (in most cases) for the tools with a rake angle γ of 5° and a helix angle λs of 50°. The results provided both theoretical and practical knowledge about the achievable surface roughness after rough milling using tools with different tool blade geometry. It was shown that rough milling is an effective and efficient type of machining for the AZ31B alloy.