Stem and sheath diseases caused by Sclerotium oryzae Cattaneo (SCL) and Rhizoctonia oryzae-sativae Sawada Mordue (ROS) can severely reduce rice (Oryza sativa L.) yield and grain quality. Genetic resistance is the best strategy to control them. Phenotypic selection for resistance is hampered due to a heterogeneous distribution of the inoculum in the soil that generates high environmental variability and decreases genetic gain. To have higher selection accuracy it is necessary to develop phenotyping methods with high repeatability and discriminative power. Comparison of greenhouse methods have been reported for Rhizoctonia solani Kühn, a more invasive pathogen than SCL and ROS, and for SCL, but no such comparisons are reported for ROS. Our study compares five inoculation methods for SCL and ROS to identify the more discriminant and repeatable method and to apply it for high-throughput phenotyping of hundreds of rice lines. A method that uses an agar disc with growing mycelium attached to the base of stems was found to have the best balance between discrimination among genotypes and variability among replicates of the same genotype for both pathogens. This method was used in five greenhouse experiments for phenotyping resistance to SCL and ROS in a population of 641 rice advanced breeding lines. Heritabilities of resistance ranged from 0.36 to 0.71 in these experiments. These findings have a direct application in screening for resistance of rice to SCL and ROS, and in high-throughput phenotyping for mapping loci associated to disease resistance.