The most robust design in field experiments of coconut is the randomized complete block design (RCBD) with plot size of six palms.  It has now become a serious problem in finding homogenous blocks for experiments in coconut and therefore a single palm plot has been recommended by the authors.  This study found that single palm plot with covariate as pre yield data is the most suitable method to reduce the unknown variability of coconut palms based on analysis of data of seven concluded field experiments on cultural management and fertilization of coconut with plot size of six palms.  The percentage reduction of CV in single palm plot with covariate with respect to single palm plot without covariate varied from 21% to 60% irrespectively location, period and type of the experiment.  The percentage reduction of CV in single palm plot with covariate with respect to six palm plots without covariate was 40% to 71%. Thus single palm plot is recommended for field experiments in coconut except under special circumstance such as for coconut based farming system trails (CBFS).The methodology developed to select single palm from the concluded experiments with plot size more than one can easily be applied for similar studies in other tree crops.