With the development of science and technology, electricity theft tends to be more intelligent and covert. A new kind of electricity theft based on the state estimation, namely false data injection, is analyzed. First, a standardL0-norm optimization model is proposed based on the classic false data injection model. The effects of variable factors on the solution accuracy are eliminated with this standard model. Furthermore, in order to increase the cost and difficulty of the electricity theft, the placement of PMUs based on the vulnerability index is proposed. Finally, the standard L0-norm optimization model and the countermeasure are tested on the IEEE Power Flow Test Cases. The results show that the proposed standard L0-norm optimization model can obtain more accurate false data injection vector, and setting the PMUs based on the vulnerability index can increase the electricity theft cost of illegal users effectively.