The present disclosure is directed to systems and methods for performing load balancing and message routing by a device intermediary to a plurality of short message peer to peer (SMPP) clients and a plurality of SMPP servers. The device can receive a request from an SMPP client to establish an SMPP session, replace a first sequence identifier in the request with a second sequence identifier generated by the device, and store a mapping of the second sequence identifier to the first sequence identifier. The device can select an SMPP server to forward the request with the second sequence identifier and receive a response from the SMPP server with the second sequence identifier. The device can identify, from the mapping, the first sequence identifier and the connection to the SMPP client using the second sequence identifier to forward the SMPP response with the first sequence identifier.