Thermometers in laboratory environment and industrial applications are often subject to extraneous, usually unwanted and uncontrolled magnetic fields. Magnetic field influence can be minimized, but cannot be fully cancelled out. Even more, in most cases, there is no awareness of the existence of magnetic fields, let alone their effect on measurement instrumentation. In the past, sensitivity to high dc magnetic fields has been investigated in cryogenics and at high temperatures. More recently, the magnetic effect on weak dc magnetic fields was presented. The goal of this paper was to analyse and empirically and experimentally prove the magnetic sensitivity of thermocouples exposed to low magnetic fields: both dc and ac. Precision and uniform alternating and direct magnetic flux densities were generated by means of permanent magnets and power amplifiers with air-cored coils. The magnetic effect on ferromagnetic and non-ferromagnetic thermocouples at liquid-nitrogen-boiling point (−196 °C), ice point (0 °C), in water (17 °C) and at melting point of gallium fixed point cell (29.7646 °C) was investigated. Magnetic-field-dependent temperature errors of up to 700 mK (at 5.3 mT: dc) and up to 1 °C (at 10 mT: ac 50 Hz magnetic fields) were detected. From the results, it can be concluded that, ideally for temperature measurements of the highest accuracy in the above-cryogenic temperature range, magnetic sensitivity should be estimated and taken into account either as the correction of an error and/or as an additional source of measurement uncertainty. Special consideration should be given to thermocouple orientation relative to the magnetic field direction, influence of metal enclosures and magnetization effects on ferromagnetic components of thermocouples.