Statement of problem: Few studies have compared the retention of resin-based and zinc-oxide noneugenol provisional cements. Crown retention testing is difficult and variable; therefore, a simpler method of testing by using flexural strength is indicated. Purpose: The purpose of this study was to measure the retention of base metal alloy castings to dentin provided by 8 provisional cements (3 resin-based and 5 zinc oxide) and correlate the retention to their flexural strength. Material and methods: Flexural strength specimens (2.5 × 2.5 × 22 mm) were made of each cement. The specimens were placed in a 3-point bending testing mode and loaded at 1 mm/min until failure. Ten extracted teeth were milled to a standardized complete crown preparation, fitted with Rexillium III castings, and cemented with each provisional cement. The specimens were subjected to a tensile load (1 mm/min) until failure in a universal testing machine. The flexural strength for each cement was calculated in MPa and correlated to the retention. A 1-way ANOVA and Tukey HSD post hoc analysis were used to determine significant intergroup differences (α=.05). Linear regression was used to correlate flexural strength and crown retention (α=.05). Results: Significant differences were found in the flexural strengths and retention provided by the various cements. Flexure strength was correlated with cement retention for resin-based cements (r=0.998) but not zinc-oxide noneugenol cements (r=0.058). Conclusions: Based on a 20-degree preparation, stronger cements provide increased retention. Therefore, the desired amount of retention should be based on both the cement and a clinical evaluation of the preparation. (J Prosthet Dent 2007;98:455-460). © 2007 The Editorial Council of the Journal of Prosthetic Dentistry.