Paste and mortar specimens were manufactured by using ordinary portland cement (OPC), C3A-free portland cement, slag cement and pozzolan cement. A carbonaceous or siliceous filler (10% by cement weight) was blended with each of the above portland cements. Limestone or quartz sands were used for mortar mixtures. Four different water-cement ratios (w/c) were adopted: 0.55, 0.50, 0.45, and 0.40. After a 28-day wet curing, paste and mortar specimens were immersed in magnesium sulfate aqueous solutions with a sulfate concentration of 350, 750, and 3000 mg/l, corresponding to chemically aggressive exposures 5a, 5b, and 5c respectively, according to the European Norms (ENV 206). The deterioration of cement paste specimens was studied by X-ray diffraction analysis to detect ettringite and/or thaumasite formation in relationship with the visual observation of sulfate attack. The deterioration of mortar specimens was studied by measuring elastic modulus and compressive strength at different periods of aggressive exposure (from 1 month to 5 years). After 5 years of exposure to the sulfate attack, paste and mortar specimens with slag and pozzolan cements were undamaged independent of the sulfate concentration, sand type, and w/c. On the other hand, paste and mortar specimens with blended limestone-portland cements showed surface damage when exposed to the 3000 mg/l of sulfate aqueous solution. However, they did not show loss of either compressive strength or elastic modulus in the 5 years of sulfate exposure. The surface damage was mitigated when OPC was replaced by the C3A-free portland cement and completely eliminated when this cement was blended with a siliceous filler rather than with a limestone case. Thaumasite and ettringite are responsible for the surface attack. The amount of thaumasite was a little higher in the presence of blended limestone portland cement.