ESPs are important in oil production, but their efficiency drops a lot when they handle viscous fluids. When these conditions exist, the usual water-based performance curves are not reliable and can result in wrong predictions of pump head and flow rate. This chapter demonstrates how to identify and optimize the parameters of KSB (Kreiselpumpen-Lexiko, KSB, 1989) and Gülich (Centrifugal pumps, 2008) models to enhance the accuracy of predicting ESP performance in viscous conditions. Experimental data from a TE2700 ESP working with viscosities up to 520 cP and speeds from 1800 to 3500 rpm were used to calculate and optimize the correction factors for head and flow using Nelder–Mead and Levenberg–Marquardt algorithms. The optimized KSB model performed well, reducing errors significantly and achieving R2 values above 0.98 and MAPE below 13%. The attempt to linearize nonlinear models was not successful because the models were too complex. In addition, a multi-linear regression model was built which is simple but does not generalize well. The study introduces a detailed computational system to enhance ESP modeling in viscous conditions and suggests future research topics such as multiphase flow modeling and AI-based regression.