Low-density parity check (LDPC) codes have been gradually used in Wireless Sensor Networks (WSNs) owing to their strong error-correcting ability, low coding complexity, and high decoding efficiency. The shortcomings of the IEEE802.16e standard for LDPC coding algorithms are the fixed code length and code rates. These requirements are difficult to meet because LDPC code shortening and puncturing methods are currently underdeveloped. In this paper, we present an efficient adaptive LDPC coding method based on confidence criteria. First, a log likelihood ratio-based belief propagation (LLR-BP) algorithm iteratively generates a confidence value that is used to determine the priority ranking of shortening and puncturing positions across an information sequence. At transmission, the shortening position of the information sequence is set to 0 whenever the code length is less than one LDPC code word. Coding can be completed rapidly using the information sequence along with the original information bits. Faster coding can also be achieved by using the original information when the current code rate is lower than the rate demanded. The coding sequence can be sent after deleting the puncturing position corresponding to the verified bit. Knowing that this procedure has been applied enables BP decoding to be performed at the receiving end. Simulation results show that the method can adaptively adjust the code length and code rate and effectively improve the performance of LDPC coding algorithms. Thus, the anti-jamming ability of WSNs is enhanced.