Knowledge convergence is an important means of innovation. The study aims to explore how knowledge convergence influences innovation performance at an organizational level. Furthermore, we address the moderating role of network relational embeddedness on the innovation deriving from knowledge convergence. Our empirical analyses adopting negative binomial regression models employ patent counts and patent citations from the nanotechnology field. The findings reveal that the scientific intensity in the convergence between scientific knowledge and technological knowledge has an inverted U-shaped influence on innovation performance and that this association is flattened in organizations with high network relational diversity. Also, we find that the technological scope in convergence of technological knowledge self has an inverted U-shaped influence on innovation performance and that this association is steepened in organizations with high network relational strength. Our findings add understandings of knowledge convergence on organization innovation and also have important practical and political implications.