Studies on artificial laboratory selections with abamectin, cross-resistance relationships, and stability of resistance were carried out with Tetranychus urticae Koch to provide basic information for an abamectin resistance management program. Selections for resistance and susceptibility to abamectin were performed in a population of T. urticae, collected from a commercial strawberry field in the State of São Paulo, Brazil. After five selections for resistance and five selections for susceptibility, susceptible (S) and resistant (R) strains of T. urticae to abamectin were obtained. The resistance ratio (R/S) at the LC50 reached 342-fold values. The toxicity of eight acaricides was evaluated in the R and S strains, observing significant differences (at LC50) between R and S strains for milbemectin, fenpropathrin and chlorfenapyr. Significant correlation was detected between the LC50s of abamectin and milbemectin, indicating cross-resistance between these acaricides. No cross-resistance was detected for the acaricides fenpyroximate, cyhexatin, propargite and dimethoate. The stability of abamectin resistance was also studied under laboratory conditions. Abamectin resistance was unstable in the absence of selection pressure. For all studied populations (with 75, 50 and 25% of initial frequency of resistant mites), the percentage of resistant mites decreased to levels equal or lower than 15% in six months. The results indicate that milbemectin should be avoided for managing abamectin resistance in T. urticae.