The supply of sulfur (S) fertilizers, such as phosphogypsum, to new agricultural frontiers has been hindered by the high cost of freight. However, this problem could be resolved by utilizing deposits of rock in the affected regions. Accordingly, a greenhouse study was designed to evaluate the effect of S source and rate on soybean yield. Five S rates (0, 50, 100, and 200 mg kg−1) and five S sources (S-phosphogypsum, S-Niquelândia rock, S-Araripina rock, S-Grajaú rock, and S-Sulfurgran [90% S0 + 10% bentonite]) were applied to soybean grown in two soil types (Typic Ultisol and Typic Oxisol), which differ in clay content. Regardless of soil type, the application of P-phosphogypsum resulted in a higher grain yield. However, application of the other S sources also improved grain yield as well as total dry weight; S application, in general, improved soybean yield components (number of grains per pod, number of pods per pot, and weight of pods per pot), photosynthetic rate, chlorophyll content, and the S levels of leaves and grains as well as the available sulfate (SO42-) levels in the soil. Therefore, for soybean grown in Typic Ultisol and Typic Oxisol, different S sources effectively improve a variety of variables that ultimately improve grain yield.