AbstractThis study builds a large database from an extensive survey of existing bond strengths between ultrahigh-performance concrete (UHPC) and normal-strength concrete (NC), with total 563 specimens being collected. To make up for some factors not included in the existing tests, an additional 38 specimens were tested with the improved slant shear method to enrich the database. The four governing factors for the interface strength were identified from the test results of the total 563+38 specimens: compressive strengths of UHPC and NC materials, interface roughness, normal stress level, and casting sequence. It is highlighted that this work discusses the effect of casting sequence. To obtain an accurate prediction formula, an artificial neural network (ANN) model is constructed, where the effect of casting sequence was firstly introduced. Based on the trained ANN model, an explicit formula is presented that significantly minimizes the prediction error. A modified shear-friction formula is also proposed for the UHPC–NC interface shear strength. Compared with other calculation models for concrete interface strength, this proposed formula with multiple parameters has a better accuracy.