Data dissemination by the use of periodic broadcast is a viable scalable mechanism due to the explosion of information services in wireless networks. Many research works have exploited the data broadcast methodologies, however, upon the premise that the broadcast traffic is static on a single broadcast channel. In practice, the broadcast traffic changes dynamically and multiple broadcast channels can be utilized for efficient data dissemination. Accordingly, in this paper we devise an adaptive multichannel data-dissemination mechanism, including two components: the multichannel traffic-awareness and the deterministic balance search techniques. The multichannel traffic-awareness technique performs a cyclic estimation of dynamic access frequency distribution in response to traffic dynamics. With the measured estimation, the deterministic balance search technique adjusts channel allocation and item classification to further minimize and balance the push access and the pull response time. The experimental results show that the proposed mechanism is able to attain the result of channel partition and item classification very close to the optimum and, thus, minimize the mean access time in both push and pull channels simultaneously. Therefore, the reliability of broadcast contents is guaranteed.