Convective heat transfer coefficients for external building surfaces (hc,ext) are essential in building energy simulation (BES) to calculate convective heat gains and losses from building facades and roofs to the environment. These coefficients are complex functions of, among other factors, building geometry, building surroundings, building facade roughness, local air flow patterns and temperature differences. Previous research on hc,ext has led to a number of empirical models, many of which are implemented in BES programs. This paper first provides an extensive overview of such models for hc,ext calculation implemented in BES programs together with the corresponding assumptions. Next, the factors taken into account by each model are listed, in order to clarify model capabilities and deficiencies. Finally, the uncertainty related to the use of these models is discussed by means of a case study, where the use of different models shows deviations up to ±30% in the yearly cooling energy demand (in relation to the average result) and ±14% in the hourly peak cooling energy demand of an isolated, well-insulated building, while deviations in yearly heating energy demand are around ±6%. The paper concludes that each model has a specific range of application, which is identified in this review paper. It also concludes that there is considerable uncertainty in the prediction of hc,ext, which can be transferred to the BES results. This large uncertainty highlights the importance of using an appropriate convection model for simulations of a specific building, certainly for calculating cooling demands and related important performance indicators such as indoor temperatures, indoor relatively humidity, thermal comfort, etc.