For the first time a comparative analysis of series over the southern territory of the Russian Far East monthly sums of precipitation for the warm period (April - October) from 1949 to 1996 is carried out with help of the one-dimensional, nonparametric Kendall’s test for monotonous trends and method one-dimensional rank Theil’s regression. The estimation of the linear trends by the least squares method leads to overestimate of parameter of an inclination linear trend by absolute meaning approximately of 10%. The nonparametric approach is more realistic takes into account asymmetry of an empirical cumulative distribution function of the monthly precipitation totals. The analysis has shown that on a hum noise of considerable interannual oscillations the separate significant linear trends exist. The Kendall’s test has revealed only 2% of stations with positive and 2% of stations with negative trend. Theil’s test has revealed 7% of stations with negative trend and 5% with positive trend. However all significant trends, which were found with help of Kendall’s test, are part of Theil’s test results. Scattering of the founded trends in space and in time does not allow to speak about existence steady areas with significant trends of the monthly precipitation totals. The large spatial-temporary scattering of the detected trends is an evidence that large-scale regions with steady trends are not existing for long time (more then one month). In common, analysis showed that for the study period (1949-1996) the significant changes of the precipitation regime were not registered. So, we can believe that the large interannual changing of monthly precipitation totals is a main part of common variability and global climatic trends endow only small part to common variability. Also the given fact have been noted in works N. Tase, S. Nakagawa (1990), J. Matsumoto, H. Yanagimachi (1991), Th. Gan (1995), N. Dashko, S. Varlamov, Y. Ho Han, Y. Seup Kim (1997).