Abstract: Knowing the X-ray spectrum is important to dual energy computed tomography(DECT)reconstruction. Whereas the spectrum is not always available in practice. In addition, the reconstructed image of DECT is extremely sensitive to noise which demands special noise suppression strategy in reconstruction algorithm design. In this paper, we develop an iterative algorithm with simultaneously updating the spectrum and the image of DECT, which did not need the knowledge of the spectrum in advance. Namely, spectrum estimation and iterative reconstruction are incorporated in an iterative framework. The estimated spectrum and the reconstructed images are obtained simultaneously with the proposed algorithm. Two sets of simulation experiment with different phantoms were adopted for evaluation. In comparison with the initially estimated spectrum,experimental results validated that the proposed method can provide the accurate estimated spectrum. Besides, the quality of the reconstructed density images has been improved by this work at the same time.