Image reconstruction of electrical capacitance tomography (ECT) is an ill-posed inverse problem. The simulated annealing (SA) algorithm is suitable for the solution of the ECT inverse problem. However, selection of related parameters of the SA algorithm will influence the reconstruction performance of ECT. We present an ECT image reconstruction method based on the adaptive simulated annealing (ASA) algorithm. We adopt the bat algorithm in the new solution generation strategy of the ASA algorithm. Moreover, the definition of the energy function introduces the sparsity of the reconstructed image. As a result, an adaptive annealing strategy is proposed to select the appropriate annealing rate. We made reconstructed image comparisons among linear back-projection, Landweber iteration, SA, and ASA algorithms via simulation and static experiment. Results show that improved reconstructed images can be obtained using the ASA algorithm.