The unknown magnetization directions in the presence of remanence have posed great challenges for interpreting magnetic data. Estimating magnetization directions based on magnetic measurements, therefore, has been an active area of research within the applied geophysics community. Despite the availability of several methods for estimating magnetization directions, quantifying the uncertainty of such estimates has remained untackled. We have investigated the use of the magnetization-clustering inversion (MCI) method for the purpose of assessing the uncertainty of the recovered magnetization directions. Specifically, we have leveraged the fact that the number of clusters that one expects to see among the magnetization directions recovered from MCI needs to be supplied by a user. We propose to implement a sequence of MCIs by assuming a series of different cluster numbers, and subsequently, to calculate the standard deviations of the recovered magnetization directions at each location in a model as a practical way of quantifying the uncertainty of the estimated magnetization directions. We have developed two different methods for the calculations of the standard deviations, and have also investigated the maximum number of clusters that one needs to consider to reliably assess the uncertainty. After the proof-of-concept study on a synthetic data set, we applied our methods to a field data set from an iron-oxide-copper-gold deposit exploration in the Carajás Mineral Province, Brazil. The high-confidence zones that correspond to low-uncertainty zones indicate a high spatial correspondence with the mineralization zones inferred from the drillholes and geology.