Horizontally curved composite box girder bridges are used in interchanges of modern highway systems. This type of structure has created design problems in estimating its live load. North Americans Codes of Practice recommends some analytical methods for design of such curved bridges. However, practical requirements arising during the design process necessitate a simple design method. On the basis of the literature review, such load distribution factors due to CHBDC truck loading are as yet unavailable. An extensive parametric study, using the finite-element modelling, was conducted, in which 225 prototype bridges were analysed to evaluate their shear distribution factors when subjected to CHBDC truck loading conditions. The parameters considered were number of steel boxes, number of lanes, span length, and span-to-radius curvature ratio. Based on the data generated, empirical expressions for shear distribution factors were deduced. An alternative to the developed expressions were introduced using the Artificial Neural Network (ANN) application.