Abstract Real data processed in practical applications of mathematical models are frequently presented as a (additive or multiplicative) mixture of a deterministic - usually unknown - numeric value and some kind of noise. The noise can be considered for a random variable and treated by classical probabilistic methods. But in many cases the stochastic description of the noise is not fully adequate to the nature of the considered data, and a fuzzy representation of their non-deterministic component does much mòre correspond to their structure. Numerical data entering applied models are mathematically processed, at least by means of elementary arithmetic operations. In case of fuzzy-contaminated data such processing is limited by their algebraic properties which are not identical with the classical ones known for deterministic numbers. In this contribution we present a brief survey of the algebraic properties of such operations applicable to the fuzzy data.