In allusion to diificult to obtain the global optimal solutions due to the problems of much variables, high dimensions and sophisticated conditions in solving transmission network planning, bat algorithm is applied. The modified bat algorithm is proposed to solve low precision and local optimal solution of original algorithm through optimization of the fitness optimal individuals of bat first n% fitness by chaotic sequences generated by self-logical function. Power network planning model is designed and constructed which takes line investment cost, overload cost of normal operation, transmission corridor cost, and N and N?1 overload penalty cost constraints as objectives. By testing node IEEE-18 and Southern Brasilian 46-bus system with Matlab, the results show that the new method is feasible and efficient to solve transmission model of large dimension power transmission grid.