Stalk lodging in maize (Zea mays L.) causes yield losses estimated to range from 5 to 20% annually worldwide. Selection for rind penetrometer resistance (RPR) has proven useful in enhancing germplasm for stalk strength, and therefore improving stalk lodging resistance. We conducted quantitative trait locus (QTL) analysis for RPR in four F2:3 populations. The populations were constructed by means of combinations of MoSCSSS‐High (selection for high RPR), MoSCSSS‐Low (selection for low RPR), MoSQB‐Low (selection for low stalk crushing strength), inbred line Mo47, and inbred line B73. Individuals in each population were genotyped for simple sequence repeat (SSR) or restriction fragment length polymorphism (RFLP) markers, and data were collected for RPR over multiple locations and replications. Means combined over environments were used as trait data for composite interval mapping by QTL Cartographer. Eight, 10, eight, and nine single‐effect QTL and four, two, zero, and five epistatic interactions were detected for RPR in the four populations. Multilocus models, including the single‐effect QTL and epistatic interactions, accounted for 33.4, 44.7, 48.4, and 58.7% of the total phenotypic variation. These data clearly indicate the complex nature of stalk strength. One chromosomal region contained a QTL from all four populations, while two QTL were in common among three of the four populations and five QTL were in common between two populations. Candidate genes that overlap QTL confidence intervals include those involved in lignin synthesis, the phenylpropanoid pathway, and the timing of vegetative phase change.