In order to optimally control the interior permanent magnet synchronous machine (IPMSM) drives, maximum torque per ampere (MTPA) control has been widely applied to various applications, e.g., electric vehicle, high-speed rail, etc. After around 30 years’ development, the MTPA control schemes which calculate the MTPA points online based on the machine parameters became one kind of the most important and well-adopted MTPA control methods. However, according to the latest researches, since the most of these MTPA control schemes ignore the terms of the partial differentiation of machine parameters with respect to current angle, therefore, even substitutes accurate machine parameters into these MTPA control schemes, these traditional online MTPA point calculation schemes will still result in large MTPA control errors and there is no research on the observation of these differentiation terms has been reported yet. Based on the above findings, this project will innovatively develop novel observers to observe the partial differentiation of machine parameters with respect to current angle terms online and comprehensively study the effects of these differentiation terms on MTPA point calculation. Further, the project will develop robust MTPA control schemes with high MTPA control accuracy but insensitive to machine parameter errors and the developed MTPA control schemes will also be extended to flux weakening region.