Crowdsourcing systems which rely on a great deal of crowdworkers to perform large quantities of microtasks, have been leveraged in a variety of applications. There are two factors affecting the productive output of each crowdworker. One is skill level, which is private information to each crowdworker, and another is her variable expended effort. In this paper, we construct and analyze a total-Ability-balanced team based incentive mechanism ABT, which can stimulate the strategic crowdworkers to truthfully report their ability levels, and according to crowdworkers' ability levels, form the competing teams. Specifically, a crowdworker with a certain skill level, is askedto choose a specificskill level (i.e., denoted as an ability threshold), and a basic payment scheme is designed to incentivize the crowdworker to truthfully report her ability level. Then, according to the chosen ability thresholds, crowdworkers are organized into total ability balanced teams to earn extra team bonus, which can further motivate crowdworkers to exert more efforts. Compared to team formation process where workers are randomly assigned to the same-scale teams and the pay per task model, our scheme ABT can improve the work efficiency.