It is increasingly popular to optimize geochemical reaction models to fit lab or field data. Performing these optimizations is often technically difficult due to the highly correlated nature of the parameters and the common presence of local minima in the optimization error surface. However, there are simple techniques for recognizing when minima will occur and ways to remove many of minima from the optimization problems. If the cause of the minimum is known and the sum of the square weighted residual (SSWR) is the measurement of error, then adding terms to the SSWR and restating the optimizations goals may remove the minima. The SSWR term might also need to be redefined to include quantifying the error in more dimensions. Also, one of the best tools for understanding where minima might occur is through creating phase diagrams of the parameter space so that you can understand phase equilibria and where parameters may change or be static.