Scientific planning the capacity of distributed generation (DG) will contribute to the economical operation, reliability and power quality of the distribution network. A fast determination method based on conic programming (CP) is proposed for solving DG allocation problem in this paper, which focuses on studying the transformation method from an initial optimal capacity model to a conic one. Then a conic optimal model is integrated with full consideration on the standard form of CP, and consequently realizes consolidated solving for the optimization and power flow calculation rapidly. Additionally, the proposed method is applied to IEEE 33-feeder test case and results are compared with traditional algorithms. Numerical studies are validating the availability of the conic transformation model as well as representing of the effectiveness and out-performance of the CP algorithm. ©, 2014, Chinese Machine Press. All right reserved.