The approximated nonlinear least squares (ALS) method has been used for adjusting unknown universal constants in the complex simulation code, which is very time-consuming to execute. The ALS tunes or calibrates the computer code by minimizing the squared difference between real observations and computer output using a surrogate such as a Gaussian process model. A potential drawback of the ALS is that it does not take the uncertainty in the approximation of the simulation by a surrogate model into account. The calibration result is too dependent on selection of a surrogate model. To address these problems, we consider a simple ensemble method that averages the respective estimates from four different models. A total of three test functions in different conditions are examined for a comparative analysis. Based on the test function study, we find that the ensemble method by four models provides better results than the ALS method. We also review some calibration methods including a generalized ALS, an iterative version of ALS, and likelihood-based method. We provide a brief discussion for comparison and future direction.