The general, 3-parameter Pareto distribution has been and is being used for modeling distributions of interest in reliability and commerce. Charek, Moore and Coleman (1988) compared minimum distance estimation with best linear unbiased estimation for the location and scale parameters of the 3-parameter Pareto distribution. However he Charek et al. paper and the Kulldorff and Vanninan (1973) paper which derived the best linear unbiased estimates for the general 3-parameter Pareto distribution both assumed that the shape parameter was known and specified. Our paper shows that the estimates of the location and scale parameters are sensitive to the miss-specification of the shape parameter. in order to correct for this possible miss-specification, two least squares type estimation procedures are proposed for estimating all 3 parameters. It is also indicated in this paper that in many cases the likelihood function surface is frustratingly flat especially in the direction of the shape parameter and this flatness makes maximum likelihood estimation unstable. A Monte Carlo study of the properties of the 2 least squares type estimators is included and the estimators are compared for bias and mean square error.