In this paper, we propose an efficient method for generating two types of novel optimized long binary spreading sequences (OLBSS) with improved autocorrelation function (ACF) properties. The first type is constructed from concatenated short binary subsequences belonging to the same code family, such as Walsh Hadamard and Gold subsequences, provided that their crosscorrelation functions (CCFs) have good properties. The second category uses the same subsequences but which are rather interlaced. Here, the number and size of the subsequences are related to the chosen length of the final constructed long sequence and the desired performances. The realization of the OLBSSs is achieved using two different optimization techniques, namely, the genetic algorithms (GAs) and particle swarm optimization (PSO) method. The simulation results, based on MATLAB tool, have shown that the proposed long sequences, composed of Walsh-Hadamard subsequences and optimized by the GA, have better ACF properties compared to the original Gold, Weil, and random sequences of the same length.