In this study, we develop an independent method to derive global temperature trend from Microwave Sounding Unit (MSU) radiometer observations in Ch 2, made from sequential, sun- synchronous, polar-orbiting NOAA operational satellites. Also, a detailed examination of the systematic errors in these data is performed with the objective to improve this method. Partitioning these data from 75 N to 75 S into global land and ocean sets, and further subdividing them into AM and PM subsets with the help of the LECT, enables us to perform this examination. The systematic errors in the MSU Ch 2 data are mainly related to differences in the MSU instrument calibration and the Local Equatorial Crossing Times (LECT) of successive satellites. They can be removed from these data with the help of the overlapping observations made by successive satellites. Errors in the MSU Ch 2 data are also introduced by orbital drift, which is the progressive change in the LECT of a satellite. Changes of the AM and PM observation times due to orbital drift causes Ch 2 brightness temperatures from each satellite to be affected by the diurnal cycle. In addition, orbital drift alters satellite illumination by sun, and thereby the instrument calibration. These errors introduced by orbital drift cannot be eliminated objectively. However, in this study, the uncertainty in the global temperature trend resulting from the cumulative error generated by drifts of all the satellites is inferred with an indirect approach. Based on our method of analysis of the MSU Ch 2 data, we find a global temperature trend from 1980 to 1996 of 0.11 K decade -1 with an uncertainty of 0.06 K decade -1.