Walsh and Gold codes are binary valued and are widely used in wireless CDMA applications due to their ease of generation as well as their fixed power property. There are not many code families available in the binary sample space with good performance. Intra- and inter-code correlations and in turn performance might be improved if we relax the sample values of spreading codes. In this paper, we attempt to obtain multiplevalued integer codes with better performance than the currently employed binary codes. We present two methods to design such spreading codes. In the first method, multiple valued spread spectrum code solutions are generated using brute-force search method within the integer numbers space. As a second method, Kronecker product (hybrid) codes of two or more families are generated. We designed several sample codes by utilizing these techniques and compared their BER performances with the known solutions for multiuser CDMA communications. It is shown that the proposed design techniques yield spreading codes that perform comparable to or superior to the Walsh and Gold families under the same test conditions. We predict that increasing the available spreading code families with various short and long code lengths, and good performance brings more efficiency and security for the design of software radio and dense sensor systems in their future implementations.