What is the differences between predict and forecast value ?

i have a targetSeries and i wanna forecast next value.i calculate autocorrelation and find feedbackDelay in narnet. testdata is %30 of TargetSeries,traindata is %70 of targetSeries and myMape is about 3.18 of prediction but when i forecast for last %30 myMape is about 14.30. what is the differences between predict and removedelay ? 2-How can i select best FeedbackDelay?

   N = length(TargetSeries);
    zt = zscore(TargetSeries,1);
    autocorrt = nncorr( zt, zt, N-1, 'biased' );
    [ sortact FeedbackDelay] = sort(autocorrt(N:end),2,'descend');
  for i=1:length(sortact)
      if(sortact(1,i)>=0.80) 
      feedbackDelay(1,i)=FD(1,i);
      end
  end 
  feedbackDelay = sort(feedbackDelay);
  net = narnet(feedbackDelay,hiddenLayerSize,'open',trainFcn);
  %train
  [net tr Ys Es Xf Af] = train(net,inputs,targets,inputStates,layerStates);
  %test
  yPrediction = net(testdata,inputStates,layerStates);
  %MAPE
  MAPE=(sum(abs(abs(OrijinalTestNumbers-yPredictionNumbers)./OrijinalTestNumbers))/length(OrijinalTestNumbers))*100;%about 4.23
  Forecast = traindata;
  for i=1:testLength
  nets = removedelay(net2);
  [xs,xis,ais,ts] = preparets(nets,{},{},Forecast);
  ys = nets(xs,xis,ais);
  Forecast(1,end+1) = ys(1,end);
  i
  end

Thank you for helps.