The behavior of hybrid nanofluid (HNF) based on SAE40 oil has been investigated using response surface methodology (RSM) in this study. Laboratory conditions in this study are temperature T = 25-50 degrees C, SVF= 0.0625-1% and SR= 666.5-9331 s-1. The purpose of this study is to select the best-predicted model and creates optimal lubrication conditions in a cold environment (or starting car). The statistical-mathematical analysis of the performance of various modelers and their comparison, as well as the non-unification of laboratory behavior on a real scale in the past literature, were the motivations of the present work. Four different models (2FI, quadratic, cubic, and quartic models) have been evaluated. One of the innovations of this study is the investi-gation of various statistical aspects of these modeling functions and their comparison. Statistical findings show that the quartic model has double accuracy in presenting the HNF properties compared to other models. The model evaluation indicators are compared based on the R2, C.D%, C.V%, and P-value. The values of these indices are for the quartic model 0.9990, -6.337 < C.D < 6.64%, 1.399196276e-216% and 1.76%, respectively. The optimum mode of viscosity of HNF for lubrication in a cold environment (T = 31.395 degrees C, SVF = 0.063% and SR = 666.510 s-1) is suggested as 212.360 MPa.sec.