Hydrogen/deuterium exchange (HDX) coupled to protein fragmentation either in solution (by means of proteolysis) or in the gas phase (using collisional activation of protein ions) and followed by mass spectral measurements of deuterium content of individual fragments has become one of the major experimental tools to probe protein structure and dynamics. One difficulty, which often arises in the course of interpretation of HDX MS data, is a need to separate deuterium contribution to the observed isotopic patterns from that of naturally occurring isotopes. Another frequently encountered problem, especially when HDX in solution is followed by protein ion fragmentation in the gas phase, is a need to determine the deuterium content of an internal protein segment based on the measured isotopic distributions of overlapping fragments. While several algorithms were developed in the past several years to address the first problem, the second one did not enjoy as much attention. Here we report a new algorithm based on a maximum entropy principle, which is capable of extracting local exchange data form the isotope distribution of overlapping fragments, as well as subtracting the background due to the presence of natural isotopes and residual deuterium in exchange buffers. The new method is tested with several proteins and appears to generate stable solutions even under unfavorable circumstances, e.g., when the resolving power of a mass analyzer is not sufficient to avoid signal interference or when the isotopic distributions of individual fragments are complex and cannot be approximated with simple binomial distributions. The latter feature makes the algorithm particularly useful when the exchange in solution is correlated or semicorrelated, paving the way to precise structural characterization of non-native protein states in solution.