Studies on artificial laboratory selections with milbemectin, stability of milbemectin resistance and possible cross-resistance with abamectin were carried out with Tetranychus urticae Koch to provide basic information for a milbemectin resistance management program. Selections for resistance and susceptibility to milbemectin were performed in a population of T. urticae, collected from a commercial chrysanthemum field in the State of São Paulo, Brazil. After six selections for resistance and five selections for susceptibility, susceptible (S) and resistant (R) strains of T. urticae to milbemectin were obtained. The resistance ratio (R/S) at the LC(50) reached 409-fold value. The stability of milbemectin resistance was also studied under laboratory conditions, using a population with initial frequency of 75% of resistant mites. The frequencies of milbemectin resistance were evaluated monthly for a period of 7 months. In order to observe possible correlation between milbemectin and abamectin resistance, the frequencies of abamectin resistance were also evaluated for that population, during the same period. The frequency of milbemectin resistance decreased from 75 to 14.5%, while the percentage of abamectin resistant mites decreased from 57 to 9.1%, in 7 months. The frequencies of milbemectin and abamectin resistance were also evaluated in 25 field populations of T. urticae, collected from several crops in the State of São Paulo. The frequencies of milbemectin resistance varied from 4.1 to 89.5%, and of abamectin, from 7.0 to 90.5%. A positive and significant correlation was observed between the frequencies of milbemectin and abamectin resistance, indicating positive cross-resistance between these acaricides. The results indicate that abamectin should be avoided for managing milbemectin resistance in T. urticae. This is the first report on milbemectin resistance in T. urticae in Brazil.