In this study neutralized and bleached canola oil, deodorized according to the Central Composite Design consisting 30 experiments with differing levels of deodorization temperature, pressure, stripping steam rate and time parameters. Free fatty acid (FFA) content, oxidation stability index (OSI), peroxide value (PV), total polar compounds (TPC), tintometric redness, fatty acid composition of every deodorized canola oil sample was measured. Deodorization parameters were optimized using only responses with good model fit by aiming lowest free fatty acid and trans-linolenic acid (tr-C18:3) contents and highest OSI value and polyunsaturated fatty acid (PUFA) content via Response Surface Methodology. Optimum temperature, pressure, stripping steam, and time were predicted as 228.8°C, 1.4mBar, 1.25 gr/min and 80 minutes, respectively and the optimization model has predicted the FFA level to be 0.044%, OSI as 10.65h, tr-C18:3 content as 0.21% and PUFA content as 30.50%. Results of the validation experiments at this optimum point were close to the those predicted and the relative differences between predicted and validation results were within the variance coefficients of each model. Temperature and time of deodorization were found significantly effective on trans isomerization of linolenic acid on ANOVA, therefore the reaction rate constants of tr-C18:3 formation and cis-linolenic acid (cis-C18:3) degradation were calculated together with Arrhenius’ equation constants using graphical method. Cis-C18:3 degradation rate was higher than that of tr-C18:3 formation showing; besides isomerization, different decomposition mechanisms took place for possibly not only for linolenic acid but also for all polyunsaturated fatty acids of canola oil during deodorization.