The article is dedicated to the investigation of adaptive algorithms of band rejection filters, which ensure a top efficiency of suppression of correlated (passive) noise in conditions of a priori parametric uncertainty during wobbling of the repetition period. The purpose of the article is the analysis of the stability of algorithms for calculation of weight coefficients of rejection filters and development of recommendations for minimizing the sensitivity of these algorithms to calculation errors. A specific form of adaptive algorithms is given for which the analysis of stability of calculation of the weight coefficients of rejection filters in the wobbling of the repetition period of probing coherent pulses is carried out. Analysis of the stability of adaptive algorithms based on the energy criterion (the coefficient of improving the signalto-noise ratio), depending on the choice of their basic parameters (rows and decomposition element), makes it possible to optimize these parameters by criterion of minimum sensitivity of algorithms to mathematical calculations errors. Recommendations are given on selection of parameters of the algorithms for calculating the weight coefficients and a procedure is proposed for calculating the modified algorithms for adaptive rejection of passive noise during the wobbling of the repetition period. As a result of optimization of algorithms by the criterion of constancy of derivatives it is established that in order to ensure a minimum value of the matrix of the derivative weight coefficients under the selected number of the expansion line, the preference should be given to selection of the element (column) of the expansion under the same number.