In　electricity　market,　as　the　power　producers＇　bidding　prices　change　with　power　outputs　of　units,　the　objective　function　of　generation　scheduling　error　optimization　is　no　longer　a　linear　model,　for　which,　β-PSO　(β　distribution-Particle　Swarm　Optimization)　method　is　proposed,　using　β　distribution　random　function　instead　of　even　distribution　random　function.　Parameters　of　β-PSO　are　adjusted　dynamically　in　working　out　solution　and　iteration　process　to　improve　the　speed　and　quality　of　the　feasible　solution,　and　to　ensure　particles　to　search　best　solutions　in　the　feasible　region　during　particle　speed　updating.　An　example　shows　that,　the　method　effectively　releases　the　difficulty　of　former　PSO　algorithm　in　finding　the　feasible　solution　of　optimization　problem.