Carbon nanotubes exhibit remarkable mechanical and electronic properties and are, therefore, being regarded as a new functional material for next generation electronics. Nevertheless, several obstacles still exist for an application in industry. The control of carriers in carbon nanotubes is of critical importance prior to an industrial application in transistors. As carbon nanotubes exhibit p-type behavior under ambient conditions, it is difficult to convert them from a p- to an n-type transistor. Also, doping control is a critical issue for applying traditional CMOS technology. Here, we discuss various approaches for preparing operating carbon nanotube transistors: i) impurity doping that employs conventional and interstitial insertion of group III or V materials, ii) chemical doping that induces charge transfer between chemicals and CNTs, iii) carrier control that utilizes the work function difference between metal and CNTs, iv) electrostatic doping that controls the carrier type by using a gate bias, and v) ambipolarity that does not use chemical doping. Advantages and drawbacks of these approaches will be discussed extensively in the text.