This article presents the results of the comparison done using five methodologies for the fitting of parameters of a shifted lognormal distribution. The methods considered in this research are: the method of moments, the method of local maximum likelihood, a method proposed by Nagatsuka and Balakrishnan, a modification of this latter method, and a new method proposed in this article. The method of local maximum likelihood provided estimators with good properties when the sample size is not too small and the method did not fail. On the other hand, the modified method of Nagatsuka and Balakrishnan exhibited the best performance from the point of view of the mean squared error. The new method proposed in this work exhibited the best performance from the point of view of a measure of goodness of fit.